[{"_1":2,"_3":-5,"_4":-5},"loaderData",{"_5":6,"_7":8,"_9":10,"_11":12},"actionData","errors","root",{"_3017":3018,"_3019":3020,"_3021":3022},"routes/_public/_layout",{"_3015":3016},"routes/_public/mcp/connectors/~namespace/~slug/_route",{"_141":142,"_15":-5,"_143":144,"_145":146,"_147":146,"_148":-5,"_149":150,"_151":152,"_153":152,"_154":152,"_16":155,"_156":-5,"_157":158,"_159":160,"_161":162,"_163":164,"_18":165,"_166":167,"_168":169,"_170":171,"_172":-5,"_173":174,"_175":176},"routes/_public/mcp/connectors/~namespace/~slug/discussions/_route",{"_13":14,"_15":-5,"_16":17,"_18":19},"comments",[],"currentUserId","meta",[28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47],"remote",{"_20":21,"_22":23,"_24":25,"_26":27},"id",8079778,"namespace","ai.withhuman.app","slug","with-human","title","withHuman",{"_26":140},{"_107":114,"_68":139},{"_134":72,"_135":136,"_137":138},{"_107":116,"_123":133},{"_107":114,"_123":132},{"_107":112,"_123":131},{"_107":110,"_123":130},{"_107":108,"_123":129},{"_107":127,"_123":128},{"_107":72,"_123":126},{"_107":81,"_123":125},{"_107":122,"_123":124},{"_107":120,"_68":121},{"_107":118,"_68":119},{"_107":116,"_68":117},{"_107":114,"_68":115},{"_107":112,"_68":113},{"_107":110,"_68":111},{"_107":108,"_68":109},{"_48":49},"script:ld+json",{"_50":51,"_52":53},"@context","https://schema.org","@graph",[54,55],{"_50":51,"_56":83,"_84":85,"_86":87,"_88":89,"_68":81,"_90":91,"_92":80},{"_50":51,"_56":57,"_58":59},"@type","BreadcrumbList","itemListElement",[60,61,62,63,64],{"_56":65,"_66":80,"_68":81,"_70":82},{"_56":65,"_66":77,"_68":78,"_70":79},{"_56":65,"_66":74,"_68":75,"_70":76},{"_56":65,"_66":72,"_68":27,"_70":73},{"_56":65,"_66":67,"_68":69,"_70":71},"ListItem","item","https://glama.ai/mcp/connectors/ai.withhuman.app/with-human/discussions","name","Discussions","position",5,"https://glama.ai/mcp/connectors/ai.withhuman.app/with-human",4,"https://glama.ai/mcp/connectors","MCP Connectors",3,"https://glama.ai/mcp","MCP",2,"https://glama.ai/","Glama",1,"Organization","founder",{"_56":104,"_68":105,"_92":106},"knowsAbout",[98,99,100,101,102,103],"logo","https://glama.ai/logo.png","sameAs",[93,94,95,96,97],"url","https://github.com/glama-ai","https://discord.gg/C3eCXhYWtJ","https://reddit.com/r/glama","https://x.com/glama_ai","https://www.linkedin.com/company/glama-ai/","Model Context Protocol","MCP server registry","MCP Gateway","Tool Definition Quality Score","AI developer tools","Large language model tool calling","Person","Frank Fiegel","https://github.com/punkpeye","content","1200","twitter:image:width","630","twitter:image:height","https://glama.ai/mcp/connectors/ai.withhuman.app/with-human/og-image","twitter:image","View discussions and comments for withHuman. Share your thoughts, ask questions, and engage with the community.","twitter:description","Discussions | withHuman","twitter:title","@glama_ai","twitter:site","summary_large_image","twitter:card","en_US","property","og:locale","og:site_name","og:url","website","og:type","og:image:width","og:image:height","og:image","og:description","og:title","href","rel","canonical","tagName","link","description","Discussions | withHuman | Glama","claimOptions",{"_3003":-5,"_3004":3005,"_3006":-5},"discoveredTools",[790,791,792,793,794,795,796,797,798,799,800,801,802,803,804,805,806,807,808,809,810,811,812,813,814,815,816,817,818,819,820,821,822,823,824,825,826,827,828,829,830,831,832,833,834,835,836,837,838,839,840,841,842,843,844,845,846,847,848,849,850,851,852,853,854,855,856,857,858,859,860,861,862,863,864,865,866,867,868,869,870,871,872,873],"discussionCommentCount",0,"favoriteCount","githubUser","glamaJsonSync",{"_787":788,"_742":789},"isFavorited",false,"isMaintainer","isSuperAdmin",[751,752,753,754,755,756,757,758,759,760,761,762,763,764,765,766,767,768,769,770],"navWarnings","publisherProfile",{"_719":720,"_721":722,"_723":531,"_724":725,"_726":727},"relatedMcpConnectors",[],"relatedMcpServers",[537,538,539,540],"relatedTestProfileStatusByRemoteServer",{},{"_494":495,"_496":497,"_498":499,"_500":501,"_502":503,"_504":447,"_505":506,"_507":508,"_139":509,"_510":-5,"_511":152,"_512":447,"_20":21,"_513":152,"_514":447,"_515":516,"_517":518,"_519":-5,"_520":-5,"_68":521,"_22":23,"_522":523,"_524":146,"_24":25,"_525":526,"_527":-5,"_26":27,"_528":529,"_450":530,"_92":531},"serverQualityScore",{"_455":456,"_457":458,"_459":460,"_461":462,"_463":73,"_464":465,"_466":467,"_468":73,"_469":73,"_470":471,"_472":73,"_473":474,"_475":467,"_476":477,"_478":479,"_480":479,"_481":82},"sponsoredCard",{"_440":441,"_442":443,"_444":445,"_446":447,"_448":449,"_450":451,"_452":453,"_454":-5},"toolSchemaChangelog",[177],"uptime","userProfilesByRemoteServer",{},"versions",[],{"_20":178,"_179":180,"_181":-7,"_182":183},699435,"observedAt","2026-10-15T08:26:04.110343Z","releaseVersion","toolChanges",[184,185,186,187,188,189,190,191,192,193,194,195,196,197,198,199,200,201,202,203,204,205,206,207,208,209,210,211,212,213,214,215,216,217,218,219,220,221,222,223,224,225,226,227,228,229,230,231,232,233,234,235,236,237,238,239,240,241,242,243,244,245,246,247,248,249,250,251,252,253,254,255,256,257,258,259,260,261,262,263,264,265,266,267],{"_268":438,"_270":439,"_272":273},{"_268":436,"_270":437,"_272":273},{"_268":434,"_270":435,"_272":273},{"_268":432,"_270":433,"_272":273},{"_268":430,"_270":431,"_272":273},{"_268":428,"_270":429,"_272":273},{"_268":426,"_270":427,"_272":273},{"_268":424,"_270":425,"_272":273},{"_268":422,"_270":423,"_272":273},{"_268":420,"_270":421,"_272":273},{"_268":418,"_270":419,"_272":273},{"_268":416,"_270":417,"_272":273},{"_268":414,"_270":415,"_272":273},{"_268":412,"_270":413,"_272":273},{"_268":410,"_270":411,"_272":273},{"_268":408,"_270":409,"_272":273},{"_268":406,"_270":407,"_272":273},{"_268":404,"_270":405,"_272":273},{"_268":402,"_270":403,"_272":273},{"_268":400,"_270":401,"_272":273},{"_268":398,"_270":399,"_272":273},{"_268":396,"_270":397,"_272":273},{"_268":394,"_270":395,"_272":273},{"_268":392,"_270":393,"_272":273},{"_268":390,"_270":391,"_272":273},{"_268":388,"_270":389,"_272":273},{"_268":386,"_270":387,"_272":273},{"_268":384,"_270":385,"_272":273},{"_268":382,"_270":383,"_272":273},{"_268":380,"_270":381,"_272":273},{"_268":378,"_270":379,"_272":273},{"_268":376,"_270":377,"_272":273},{"_268":374,"_270":375,"_272":273},{"_268":372,"_270":373,"_272":273},{"_268":370,"_270":371,"_272":273},{"_268":368,"_270":369,"_272":273},{"_268":366,"_270":367,"_272":273},{"_268":364,"_270":365,"_272":273},{"_268":362,"_270":363,"_272":273},{"_268":360,"_270":361,"_272":273},{"_268":358,"_270":359,"_272":273},{"_268":356,"_270":357,"_272":273},{"_268":354,"_270":355,"_272":273},{"_268":352,"_270":353,"_272":273},{"_268":350,"_270":351,"_272":273},{"_268":348,"_270":349,"_272":273},{"_268":346,"_270":347,"_272":273},{"_268":344,"_270":345,"_272":273},{"_268":342,"_270":343,"_272":273},{"_268":340,"_270":341,"_272":273},{"_268":338,"_270":339,"_272":273},{"_268":336,"_270":337,"_272":273},{"_268":334,"_270":335,"_272":273},{"_268":332,"_270":333,"_272":273},{"_268":330,"_270":331,"_272":273},{"_268":328,"_270":329,"_272":273},{"_268":326,"_270":327,"_272":273},{"_268":324,"_270":325,"_272":273},{"_268":322,"_270":323,"_272":273},{"_268":320,"_270":321,"_272":273},{"_268":318,"_270":319,"_272":273},{"_268":316,"_270":317,"_272":273},{"_268":314,"_270":315,"_272":273},{"_268":312,"_270":313,"_272":273},{"_268":310,"_270":311,"_272":273},{"_268":308,"_270":309,"_272":273},{"_268":306,"_270":307,"_272":273},{"_268":304,"_270":305,"_272":273},{"_268":302,"_270":303,"_272":273},{"_268":300,"_270":301,"_272":273},{"_268":298,"_270":299,"_272":273},{"_268":296,"_270":297,"_272":273},{"_268":294,"_270":295,"_272":273},{"_268":292,"_270":293,"_272":273},{"_268":290,"_270":291,"_272":273},{"_268":288,"_270":289,"_272":273},{"_268":286,"_270":287,"_272":273},{"_268":284,"_270":285,"_272":273},{"_268":282,"_270":283,"_272":273},{"_268":280,"_270":281,"_272":273},{"_268":278,"_270":279,"_272":273},{"_268":276,"_270":277,"_272":273},{"_268":274,"_270":275,"_272":273},{"_268":269,"_270":271,"_272":273},"changes",[],"toolName","update_team_escalation_policy","type","observed",[],"update_pipeline",[],"update_member",[],"update_gateway_server",[],"update_escalation_path",[],"update_agent_instance",[],"update_agent",[],"test_webhook_endpoint",[],"set_team_members",[],"set_gateway_scope_oauth_client",[],"set_gateway_scope_headers",[],"set_agent_tool_access",[],"save_webhook_endpoint",[],"rotate_webhook_endpoint_secret",[],"revoke_provisioner_token",[],"revoke_agent_credential",[],"restore_agent",[],"rename_team",[],"remove_member",[],"read_reference",[],"preview_pipeline",[],"list_webhook_endpoints",[],"list_webhook_endpoint_uses",[],"list_webhook_endpoint_revisions",[],"list_tool_definitions",[],"list_tool_catalog",[],"list_teams",[],"list_team_permission_policies",[],"list_team_escalation_policy_revisions",[],"list_roles",[],"list_pipelines",[],"list_pipeline_uses",[],"list_pipeline_revisions",[],"list_permissions",[],"list_members",[],"list_member_permission_policies",[],"list_gateway_tunnels",[],"list_gateway_servers",[],"list_gateway_server_tools",[],"list_gateway_server_revisions",[],"list_escalation_paths",[],"list_escalation_path_uses",[],"list_escalation_path_revisions",[],"list_audit_events",[],"list_audit_event_types",[],"list_agents",[],"list_agent_tool_access",[],"get_webhook_endpoint_revision",[],"get_webhook_endpoint",[],"get_tool_catalog_entry",[],"get_team_escalation_policy_revision",[],"get_team_escalation_policy",[],"get_team",[],"get_request_timeline",[],"get_pipeline_revision_references",[],"get_pipeline_revision",[],"get_pipeline",[],"get_gateway_server_revision",[],"get_gateway_server",[],"get_escalation_path_revision",[],"get_escalation_path",[],"get_audit_event",[],"get_agent",[],"disconnect_gateway_scope",[],"deactivate_gateway_server",[],"create_team",[],"create_provisioner_token",[],"create_pipeline",[],"create_gateway_server",[],"create_escalation_path",[],"create_agent_instance",[],"create_agent",[],"archive_webhook_endpoint",[],"archive_team_escalation_policy",[],"archive_team",[],"archive_pipeline",[],"archive_gateway_server",[],"archive_escalation_path",[],"archive_agent",[],"activate_webhook_endpoint_revision",[],"activate_team_escalation_policy_revision",[],"activate_pipeline_revision",[],"activate_gateway_server_revision",[],"activate_escalation_path_revision","advertiserName","Fulcra Dynamics","body","Agents share what they learn, hand off work, and keep it when you switch models. Start with 5 GB free.","destinationUrl","https://fulcradynamics.com/product/start-for-free?utm_source=glama&utm_medium=paid_sponsorship&utm_campaign=q4_2026_builder&utm_content=app-orches-v1-share","hasLogo",true,"headline","Context Every Agent Can Share","uid","cav1v6pidq","updatedAt","2026-10-07T05:30:03.592577Z","deliveryToken","coherenceJustifications",{"_463":482,"_468":483,"_484":485,"_486":487},"coherenceScore",3.3,"coherenceSummary","The withHuman server offers a rich, mostly coherent tool set with strong resource-specific naming and disambiguation. However, 84 tools is excessively large for practical agent use, risking cognitive overload despite comprehensive lifecycle coverage.","coherenceTier","B","completeness","descriptionQualityScore",3.7,"descriptionQualityTier","A","disambiguation","meanTdqs","minTdqs",3.2,"namingConsistency","overallScore",3.6,"overallTier","scoredAt","2026-10-15T09:15:37.482525Z","scoredToolCount",84,"toolCount","toolCountAppropriateness",{"_488":493,"_490":73},{"_488":492,"_490":73},"naming_consistency",{"_488":491,"_490":73},"tool_count_appropriateness",{"_488":489,"_490":82},"justification","84 tools is an extreme overload for any MCP server, far exceeding the typical 3-15 range and even the 25+ threshold. The complex domain justifies many operations, but this volume risks agent confusion and poor tool selection.","score","Tool names follow a consistent snake_case verb_noun pattern (create_agent, list_pipelines, update_pipeline, archive_webhook_endpoint, activate_*_revision). Minor deviations like save_webhook_endpoint and set_* are understandable but slightly break the create/update idiom.","Tools are largely distinguishable by resource and action (agent vs instance vs provisioner token, pipeline vs revision vs use), and descriptions clarify boundaries. However, the repeated revision/activation patterns across five resource types create some risk of misselection.","Coverage is broad and deep across agents, instances, pipelines, escalation paths, gateway servers, teams, webhooks, audit, and permissions, with full revision lifecycles. Minor gaps exist (e.g., no get_member, no list_requests), but agents can work around them via existing list tools.","apiKeyHeaderName","X-API-Key","apiKeyLocation","bearer_authentication","apiKeySearchParameterName","api_key","authType","oauth2","categoryAttributes",[535,536],"claimable","deprecatedAt","2026-10-15T10:39:43.791788Z","deprecationComment","This is a duplicate of the actual [entry](mcp/connectors/ai.withhuman/withhuman) deprecating this in favor of the original.","Set up human approval for your AI agents' tool calls.","githubRepository","glamaMetadataIsSourceOfTruth","healthy","isRegistryLinked","isVerified","lastTestedAt","2026-10-15T11:25:22.706610Z","lastVerifiedProtocolVersion","2025-11-25","linkedMcpServer","mcpServerId","ai.withhuman.app/with-human","oauthRegistration",{"_532":533,"_534":533},"recentToolCallCount","testMcpConnectionProfileId",107641,"thumbnailUpdatedAt","transportType","Streamable HTTP","b80c1njtqq","https://app.withhuman.ai/api/mcp/v1","lastOutcome","hosted","redirectUriPolicy","category:agent-orchestration","category:security-and-iam",{"_541":682,"_543":146,"_544":683,"_546":684,"_548":684,"_549":614,"_551":685,"_553":152,"_20":686,"_22":687,"_556":688,"_557":146,"_558":689,"_560":152,"_561":690,"_563":-5,"_24":691,"_480":692,"_450":693,"_452":694,"_567":152},{"_541":643,"_543":146,"_544":644,"_546":645,"_548":645,"_549":550,"_551":646,"_553":152,"_20":647,"_22":648,"_556":649,"_557":146,"_558":650,"_560":152,"_561":651,"_563":-5,"_24":652,"_480":653,"_450":654,"_452":655,"_567":152},{"_541":611,"_543":146,"_544":612,"_546":613,"_548":613,"_549":614,"_551":615,"_553":152,"_20":616,"_22":617,"_556":474,"_557":146,"_558":618,"_560":152,"_561":619,"_563":-5,"_24":620,"_480":621,"_450":622,"_452":623,"_567":152},{"_541":542,"_543":146,"_544":545,"_546":547,"_548":547,"_549":550,"_551":552,"_553":152,"_20":554,"_22":555,"_556":-5,"_557":146,"_558":559,"_560":152,"_561":562,"_563":-5,"_24":564,"_480":146,"_450":565,"_452":566,"_567":152},"addedAt","2026-06-29T01:15:15.555481Z","appCount","attributes",[607,608,609,610],"descriptionMarkdown","Browse IndustryLens's published competitive-intelligence reports and head-to-head competitor comparisons from any AI agent — real, source-backed data.","descriptionPlainText","discoverabilityScore",200,"displayName","industrylens-mcp","hasApps",56090,{"_24":587},"qualityScore","recentUsage","repository",{"_573":574,"_510":575,"_576":577,"_568":578,"_579":-5,"_580":-5,"_581":-5},"repositoryPrivate","scores",{"_568":569,"_570":462,"_571":-5,"_572":-5},"simpleIcon","industrylens-mcp-docs","q0z9zpmgio","2026-10-09T08:05:28.649218Z","isBoosted","license",100,"maintenance","quality","security","environmentVariablesJsonSchema",{"_602":603,"_604":605,"_272":606},{"_588":589,"_590":-5,"_68":564,"_591":592,"_593":594,"_595":146,"_596":597,"_598":146},"glamaJson",{"_585":586},{"_68":582,"_583":584},"npmPackage","pypiPackage","supportedPlatforms","MIT License","source","github","maintainers",[587],"IndustryLensOp","fullName","IndustryLensOp/industrylens-mcp-docs","language","owner",{"_20":600,"_601":587},"pushedAt","2026-08-17T15:55:33.000000Z","recentStargazers","spdxLicense",{"_68":582,"_92":599},"stargazers","https://spdx.org/licenses/MIT.json",3406939,"login","properties",{},"required",[],"object","author:official","category:research-and-data","environment:network","hosting:remote-capable","2026-09-16T12:00:07.153200Z",[608,641,642,609,610],"Revnuvo Company Intelligence tells AI agents what changed at a company, with evidence. It observes company websites, technologies, and DNS over time and returns timestamped, confidence-aware changes, signals, and monitoring.",300,"revnuvo-company-intelligence",99416,{"_24":634},{"_573":625,"_510":626,"_576":-5,"_568":627,"_579":-5,"_580":-5,"_581":-5},{"_568":569,"_570":624,"_571":467,"_572":569},"revnuvo-dev",9,"xk98bmad3l","2026-10-14T14:15:53.403515Z","C",{"_602":635,"_604":636,"_272":606},{"_588":628,"_590":629,"_68":620,"_591":630,"_593":631,"_595":146,"_596":632,"_598":146},{"_68":582,"_583":584},"hardeyhemy/revnuvo-dev","JavaScript",{"_20":633,"_601":634},"2026-09-16T11:18:31.000000Z",{"_68":582,"_92":599},6793590,"hardeyhemy",{"_637":638},[637],"REVNUVO_API_KEY",{"_139":639,"_272":640},"Your Revnuvo API key. Obtain a free key at https://app.revnuvo.site (100 calls/month, no credit card required). Used to authenticate to the Revnuvo MCP server.","string","category:monitoring","category:networking-and-infrastructure","2026-05-08T02:51:00.880907Z",[680,608,681],"Competitor Monitor AI - MCP server providing AI-powered tools and automation by MEOK AI Labs","Competitor Monitor AI MCP",28808,{"_24":671},2.7,{"_573":656,"_510":657,"_576":658,"_568":659,"_579":660,"_580":661,"_581":662},{"_568":569,"_570":462,"_571":624,"_572":569},"competitor-monitor-ai-mcp",11,"li9inggp98","2026-10-09T14:41:07.006320Z",{"_602":678,"_604":679,"_272":606},{"_588":672,"_590":673,"_68":652,"_591":674,"_593":675,"_595":146,"_596":676,"_598":146},{"_585":670},{"_68":582,"_583":584},{"_68":652,"_667":669},{"_666":-5,"_68":652,"_667":668},[663,664,665],"WINDOWS","LINUX","MACOS","licenseExpression","weeklyDownloads",54,7,[671],"CSOAI-ORG","CSOAI-ORG/competitor-monitor-ai-mcp","Python",{"_20":677,"_601":671},"2026-10-08T01:01:22.000000Z",{"_68":582,"_92":599},596711,{},[],"category:marketing","category:ai-and-machine-learning","2026-03-16T16:01:47.841758Z",[680,718,641,609,610],"Enables brand visibility monitoring across major AI platforms like ChatGPT, Claude, Gemini, and Perplexity. It allows users to track visibility scores, analyze competitor data, and receive actionable insights to improve AI-generated brand recommendations.","Citesurf MCP",23467,{"_24":707},3.8,{"_573":695,"_510":696,"_576":-5,"_568":697,"_579":698,"_580":-5,"_581":699},{"_568":569,"_570":624,"_571":467,"_572":569},"mcp",16,"qxas32ijwp","2026-10-13T05:35:56.969547Z",{"_602":708,"_604":709,"_272":606},{"_588":701,"_590":702,"_68":691,"_591":703,"_593":704,"_595":146,"_596":705,"_598":82},{"_68":582,"_583":584},{"_68":700,"_667":653},[665],"@citesurf/mcp","citesurf/mcp","TypeScript",{"_20":706,"_601":707},"2026-04-23T11:46:47.000000Z",{"_68":582,"_92":599},560230,"citesurf",{"_710":711,"_712":713},[710],"CITESURF_API_KEY",{"_139":717,"_272":640},"CITESURF_BASE_URL",{"_714":715,"_139":716,"_272":640},"default","https://citesurf.com","API base URL (default: https://citesurf.com)","Your API key from Dashboard > Settings","category:search","answers",{"_728":729,"_730":731,"_732":733,"_734":735,"_736":737,"_738":739},"assertedAt","2026-10-15T08:01:50.950275Z","endpointUrl","evidenceStatus","publisher-declared","reusePermissionVersion","2026-09-10","operatorName",{"_740":741,"_742":743,"_744":27},"trustCenterUrl",{"_740":750,"_742":743,"_744":750},"operatorWebsite",{"_740":748,"_742":743,"_744":749},"documentationUrl",{"_740":741,"_742":743,"_744":741},"setupRestrictions",{"_740":746,"_742":743,"_744":747},"vendorRelationship",{"_740":741,"_742":743,"_744":745},"sourceUrl","https://withhuman.ai/docs/web-app/mcp-server","status","provided","value","first-party","https://withhuman.ai/pricing","The mcp server is available on the free plan and the paid plan. There are no restrictions.","https://withhuman.ai","https://withhuman.ai/","https://withhuman.ai/security",{"_26":786},{"_107":509,"_68":139},{"_134":72,"_135":136,"_137":138},{"_107":785,"_123":133},{"_107":509,"_123":132},{"_107":112,"_123":131},{"_107":110,"_123":130},{"_107":108,"_123":129},{"_107":127,"_123":128},{"_107":72,"_123":126},{"_107":81,"_123":125},{"_107":122,"_123":124},{"_107":120,"_68":121},{"_107":118,"_68":119},{"_107":785,"_68":117},{"_107":509,"_68":115},{"_107":112,"_68":113},{"_107":110,"_68":111},{"_107":108,"_68":109},{"_48":771},{"_50":51,"_52":772},[54,773,774],{"_50":51,"_56":57,"_58":780},{"_50":51,"_56":775,"_776":777,"_139":509,"_68":27,"_778":779,"_92":72},"SoftwareApplication","applicationCategory","ServerApplication","operatingSystem","Cross-platform",[781,782,783,784],{"_56":65,"_66":80,"_68":81,"_70":82},{"_56":65,"_66":77,"_68":78,"_70":79},{"_56":65,"_66":74,"_68":75,"_70":76},{"_56":65,"_66":72,"_68":27,"_70":73},"withHuman - MCP Connector","withHuman - MCP Connector | Glama","endedAt","2026-10-15T11:25:30.923529Z","unknown",{"_874":2981,"_876":447,"_139":2982,"_878":2983,"_68":439,"_880":-5,"_556":2984,"_882":467,"_26":-5,"_450":2985},{"_874":2959,"_876":447,"_139":2960,"_878":2961,"_68":437,"_880":-5,"_556":2962,"_882":467,"_26":-5,"_450":2963},{"_874":2937,"_876":447,"_139":2938,"_878":2939,"_68":435,"_880":-5,"_556":2940,"_882":467,"_26":-5,"_450":2941},{"_874":2915,"_876":447,"_139":2916,"_878":2917,"_68":433,"_880":-5,"_556":2918,"_882":467,"_26":-5,"_450":2919},{"_874":2893,"_876":447,"_139":2894,"_878":2895,"_68":431,"_880":-5,"_556":2896,"_882":467,"_26":-5,"_450":2897},{"_874":2872,"_876":447,"_139":2873,"_878":2874,"_68":429,"_880":-5,"_556":2875,"_882":467,"_26":-5,"_450":2876},{"_874":2851,"_876":447,"_139":2852,"_878":2853,"_68":427,"_880":-5,"_556":2854,"_882":467,"_26":-5,"_450":2855},{"_874":2830,"_876":447,"_139":2831,"_878":2832,"_68":425,"_880":-5,"_556":2833,"_882":467,"_26":-5,"_450":2834},{"_874":2808,"_876":447,"_139":2809,"_878":2810,"_68":423,"_880":-5,"_556":2811,"_882":467,"_26":-5,"_450":2812},{"_874":2787,"_876":447,"_139":2788,"_878":2789,"_68":421,"_880":-5,"_556":2790,"_882":467,"_26":-5,"_450":2791},{"_874":2766,"_876":447,"_139":2767,"_878":2768,"_68":419,"_880":-5,"_556":2769,"_882":467,"_26":-5,"_450":2770},{"_874":2745,"_876":447,"_139":2746,"_878":2747,"_68":417,"_880":-5,"_556":2748,"_882":467,"_26":-5,"_450":2749},{"_874":2718,"_876":447,"_139":2719,"_878":2720,"_68":415,"_880":-5,"_556":2721,"_882":467,"_26":-5,"_450":2722},{"_874":2692,"_876":447,"_139":2693,"_878":2694,"_68":413,"_880":-5,"_556":2695,"_882":467,"_26":-5,"_450":2696},{"_874":2661,"_876":447,"_139":2662,"_878":2663,"_68":411,"_880":-5,"_556":2664,"_882":467,"_26":-5,"_450":2665},{"_874":2622,"_876":447,"_139":2623,"_878":2624,"_68":409,"_880":-5,"_556":2625,"_882":467,"_26":-5,"_450":2626},{"_874":2592,"_876":447,"_139":2593,"_878":2594,"_68":407,"_880":-5,"_556":2595,"_882":467,"_26":-5,"_450":2596},{"_874":2571,"_876":447,"_139":2572,"_878":2573,"_68":405,"_880":-5,"_556":2574,"_882":467,"_26":-5,"_450":2575},{"_874":2545,"_876":447,"_139":2546,"_878":2547,"_68":403,"_880":-5,"_556":2548,"_882":467,"_26":-5,"_450":2549},{"_874":2524,"_876":447,"_139":2525,"_878":2526,"_68":401,"_880":-5,"_556":2527,"_882":467,"_26":-5,"_450":2528},{"_874":2501,"_876":447,"_139":2502,"_878":2503,"_68":399,"_880":-5,"_556":2504,"_882":467,"_26":-5,"_450":2505},{"_874":2465,"_876":447,"_139":2466,"_878":2467,"_68":397,"_880":-5,"_556":2468,"_882":467,"_26":-5,"_450":2469},{"_874":2442,"_876":447,"_139":2443,"_878":2444,"_68":395,"_880":-5,"_556":2445,"_882":462,"_26":-5,"_450":2446},{"_874":2421,"_876":447,"_139":2422,"_878":2423,"_68":393,"_880":-5,"_556":2424,"_882":467,"_26":-5,"_450":2425},{"_874":2398,"_876":447,"_139":2399,"_878":2400,"_68":391,"_880":-5,"_556":2401,"_882":467,"_26":-5,"_450":2402},{"_874":2377,"_876":447,"_139":2378,"_878":2379,"_68":389,"_880":-5,"_556":2380,"_882":467,"_26":-5,"_450":2381},{"_874":2354,"_876":447,"_139":2355,"_878":2356,"_68":387,"_880":-5,"_556":2357,"_882":462,"_26":-5,"_450":2358},{"_874":2333,"_876":447,"_139":2334,"_878":2335,"_68":385,"_880":-5,"_556":2336,"_882":467,"_26":-5,"_450":2337},{"_874":2311,"_876":447,"_139":2312,"_878":2313,"_68":383,"_880":-5,"_556":2314,"_882":462,"_26":-5,"_450":2315},{"_874":2287,"_876":447,"_139":2288,"_878":2289,"_68":381,"_880":-5,"_556":2290,"_882":467,"_26":-5,"_450":2291},{"_874":2264,"_876":447,"_139":2265,"_878":2266,"_68":379,"_880":-5,"_556":2267,"_882":467,"_26":-5,"_450":2268},{"_874":2243,"_876":447,"_139":2244,"_878":2245,"_68":377,"_880":-5,"_556":2246,"_882":467,"_26":-5,"_450":2247},{"_874":2222,"_876":447,"_139":2223,"_878":2224,"_68":375,"_880":-5,"_556":2225,"_882":467,"_26":-5,"_450":2226},{"_874":2199,"_876":447,"_139":2200,"_878":2201,"_68":373,"_880":-5,"_556":2202,"_882":467,"_26":-5,"_450":2203},{"_874":2174,"_876":447,"_139":2175,"_878":2176,"_68":371,"_880":-5,"_556":2177,"_882":462,"_26":-5,"_450":2178},{"_874":2153,"_876":447,"_139":2154,"_878":2155,"_68":369,"_880":-5,"_556":2156,"_882":467,"_26":-5,"_450":2157},{"_874":2130,"_876":447,"_139":2131,"_878":2132,"_68":367,"_880":-5,"_556":2133,"_882":467,"_26":-5,"_450":2134},{"_874":2110,"_876":447,"_139":2111,"_878":2112,"_68":363,"_880":-5,"_556":2113,"_882":467,"_26":-5,"_450":2114},{"_874":2089,"_876":447,"_139":2090,"_878":2091,"_68":365,"_880":-5,"_556":2092,"_882":467,"_26":-5,"_450":2093},{"_874":2031,"_876":447,"_139":2032,"_878":2033,"_68":359,"_880":-5,"_556":2034,"_882":467,"_26":-5,"_450":2035},{"_874":2013,"_876":447,"_139":2014,"_878":2015,"_68":361,"_880":-5,"_556":2016,"_882":467,"_26":-5,"_450":2017},{"_874":1992,"_876":447,"_139":1993,"_878":1994,"_68":357,"_880":-5,"_556":1995,"_882":467,"_26":-5,"_450":1996},{"_874":1972,"_876":447,"_139":1973,"_878":1974,"_68":353,"_880":-5,"_556":1975,"_882":462,"_26":-5,"_450":1976},{"_874":1950,"_876":447,"_139":1951,"_878":1952,"_68":355,"_880":-5,"_556":1953,"_882":467,"_26":-5,"_450":1954},{"_874":1926,"_876":447,"_139":1927,"_878":1928,"_68":351,"_880":-5,"_556":1929,"_882":467,"_26":-5,"_450":1930},{"_874":1905,"_876":447,"_139":1906,"_878":1907,"_68":347,"_880":-5,"_556":1908,"_882":467,"_26":-5,"_450":1909},{"_874":1883,"_876":447,"_139":1884,"_878":1885,"_68":349,"_880":-5,"_556":1886,"_882":467,"_26":-5,"_450":1887},{"_874":1863,"_876":447,"_139":1864,"_878":1865,"_68":345,"_880":-5,"_556":1866,"_882":462,"_26":-5,"_450":1867},{"_874":1842,"_876":447,"_139":1843,"_878":1844,"_68":343,"_880":-5,"_556":1845,"_882":467,"_26":-5,"_450":1846},{"_874":1824,"_876":447,"_139":1825,"_878":1826,"_68":341,"_880":-5,"_556":1827,"_882":467,"_26":-5,"_450":1828},{"_874":1806,"_876":447,"_139":1807,"_878":1808,"_68":339,"_880":-5,"_556":1809,"_882":467,"_26":-5,"_450":1810},{"_874":1785,"_876":447,"_139":1786,"_878":1787,"_68":337,"_880":-5,"_556":1788,"_882":467,"_26":-5,"_450":1789},{"_874":1765,"_876":447,"_139":1766,"_878":1767,"_68":333,"_880":-5,"_556":1768,"_882":467,"_26":-5,"_450":1769},{"_874":1743,"_876":447,"_139":1744,"_878":1745,"_68":335,"_880":-5,"_556":1746,"_882":467,"_26":-5,"_450":1747},{"_874":1725,"_876":447,"_139":1726,"_878":1727,"_68":331,"_880":-5,"_556":1728,"_882":467,"_26":-5,"_450":1729},{"_874":1704,"_876":447,"_139":1705,"_878":1706,"_68":329,"_880":-5,"_556":1707,"_882":467,"_26":-5,"_450":1708},{"_874":1683,"_876":447,"_139":1684,"_878":1685,"_68":327,"_880":-5,"_556":1686,"_882":467,"_26":-5,"_450":1687},{"_874":1662,"_876":447,"_139":1663,"_878":1664,"_68":325,"_880":-5,"_556":1665,"_882":467,"_26":-5,"_450":1666},{"_874":1628,"_876":447,"_139":1629,"_878":1630,"_68":323,"_880":-5,"_556":1631,"_882":467,"_26":-5,"_450":1632},{"_874":1610,"_876":447,"_139":1611,"_878":1612,"_68":321,"_880":-5,"_556":1613,"_882":467,"_26":-5,"_450":1614},{"_874":1589,"_876":447,"_139":1590,"_878":1591,"_68":319,"_880":-5,"_556":1592,"_882":467,"_26":-5,"_450":1593},{"_874":1571,"_876":447,"_139":1572,"_878":1573,"_68":315,"_880":-5,"_556":1574,"_882":467,"_26":-5,"_450":1575},{"_874":1550,"_876":447,"_139":1551,"_878":1552,"_68":317,"_880":-5,"_556":1553,"_882":467,"_26":-5,"_450":1554},{"_874":1497,"_876":447,"_139":1498,"_878":1499,"_68":313,"_880":-5,"_556":1500,"_882":467,"_26":-5,"_450":1501},{"_874":1474,"_876":447,"_139":1475,"_878":1476,"_68":311,"_880":-5,"_556":1477,"_882":467,"_26":-5,"_450":1478},{"_874":1451,"_876":447,"_139":1452,"_878":1453,"_68":309,"_880":-5,"_556":1454,"_882":467,"_26":-5,"_450":1455},{"_874":1428,"_876":447,"_139":1429,"_878":1430,"_68":307,"_880":-5,"_556":1431,"_882":467,"_26":-5,"_450":1432},{"_874":1407,"_876":447,"_139":1408,"_878":1409,"_68":305,"_880":-5,"_556":1410,"_882":467,"_26":-5,"_450":1411},{"_874":1384,"_876":447,"_139":1385,"_878":1386,"_68":303,"_880":-5,"_556":1387,"_882":467,"_26":-5,"_450":1388},{"_874":1361,"_876":447,"_139":1362,"_878":1363,"_68":301,"_880":-5,"_556":1364,"_882":467,"_26":-5,"_450":1365},{"_874":1340,"_876":447,"_139":1341,"_878":1342,"_68":299,"_880":-5,"_556":1343,"_882":467,"_26":-5,"_450":1344},{"_874":1313,"_876":447,"_139":1314,"_878":1315,"_68":297,"_880":-5,"_556":1316,"_882":467,"_26":-5,"_450":1317},{"_874":1275,"_876":447,"_139":1276,"_878":1277,"_68":295,"_880":-5,"_556":1278,"_882":467,"_26":-5,"_450":1279},{"_874":1248,"_876":447,"_139":1249,"_878":1250,"_68":293,"_880":-5,"_556":1251,"_882":467,"_26":-5,"_450":1252},{"_874":1217,"_876":447,"_139":1218,"_878":1219,"_68":291,"_880":-5,"_556":1220,"_882":467,"_26":-5,"_450":1221},{"_874":1189,"_876":447,"_139":1190,"_878":1191,"_68":289,"_880":-5,"_556":1192,"_882":467,"_26":-5,"_450":1193},{"_874":1162,"_876":447,"_139":1163,"_878":1164,"_68":287,"_880":-5,"_556":1165,"_882":467,"_26":-5,"_450":1166},{"_874":1134,"_876":447,"_139":1135,"_878":1136,"_68":285,"_880":-5,"_556":1137,"_882":467,"_26":-5,"_450":1138},{"_874":1107,"_876":447,"_139":1108,"_878":1109,"_68":283,"_880":-5,"_556":1110,"_882":467,"_26":-5,"_450":1111},{"_874":1069,"_876":447,"_139":1070,"_878":1071,"_68":281,"_880":-5,"_556":1072,"_882":467,"_26":-5,"_450":1073},{"_874":1011,"_876":447,"_139":1012,"_878":1013,"_68":279,"_880":-5,"_556":1014,"_882":467,"_26":-5,"_450":1015},{"_874":982,"_876":447,"_139":983,"_878":984,"_68":277,"_880":-5,"_556":985,"_882":467,"_26":-5,"_450":986},{"_874":943,"_876":447,"_139":944,"_878":945,"_68":275,"_880":-5,"_556":946,"_882":467,"_26":-5,"_450":947},{"_874":875,"_876":447,"_139":877,"_878":879,"_68":271,"_880":-5,"_556":881,"_882":467,"_26":-5,"_450":883},"annotations",{"_939":152,"_940":152,"_941":152,"_942":152},"available","Save a new revision of a team's escalation policy (the whole document again, not a diff) and activate it unless activate is false. The policy orders how members are paged when a path level targets the team: levels name current members as user targets, or broadcast for everyone; every member a request reaches may still decide it. Call read_reference (escalation_paths) first.","inputSchema",{"_272":606,"_604":912,"_602":913,"_914":152},"outputSchema",{"_884":73,"_885":73,"_886":73,"_887":888,"_889":73,"_890":71,"_891":892,"_882":467,"_893":73},"tier","lgdpsadykl","behavioralTransparency","concisenessStructure","contextualCompleteness","justifications",{"_894":895,"_896":897,"_898":899,"_900":901,"_902":903,"_904":905},"parameterSemantics","purposeClarity","tdqs",4.3,"usageGuidelines","purpose_clarity",{"_488":911,"_490":71},"usage_guidelines",{"_488":910,"_490":73},"parameter_semantics",{"_488":909,"_490":73},"conciseness_structure",{"_488":908,"_490":73},"behavioral_transparency",{"_488":907,"_490":73},"contextual_completeness",{"_488":906,"_490":73},"For a complex nested-object tool with no output schema, the description covers the key agent needs: overwrite semantics, activation default, the nodes grammar pointer via read_reference, and the prerequisite. Minor gaps remain (no success/error behavior or validation notes), but annotations carry the safety profile and read_reference carries the grammar.","Annotations already declare readOnlyHint=false, idempotentHint=false, destructiveHint=false, so the safety profile is covered. The description adds genuine context beyond them: this saves a full revision rather than a diff, and activation is the default unless activate is false. It also describes the policy's runtime effect (how members get paged), which is beyond annotation scope.","Front-loaded with the core action and kept to three sentences with no filler. The middle sentence on paging behavior is dense but earns its place by explaining the domain semantics of the document being saved.","Schema description coverage is 100%, so baseline is 3, but the description adds meaning on top: it clarifies that nodes is the whole document (not a diff), that levels name current members as user targets or one broadcast target, and that activate defaults to true. This reinforces and disambiguates the schema rather than repeating it.","Provides an explicit prerequisite — 'Call read_reference (escalation_paths) first' — which is real routing guidance. It does not state when to prefer this over alternatives (e.g. activate_team_escalation_policy_revision or create vs update), so it stops short of full when/when-not coverage.","Specific verb+resource ('Save a new revision of a team's escalation policy') with scope made explicit ('the whole document again, not a diff'). The phrase 'and activate it unless activate is false' also disambiguates it from the sibling activate_team_escalation_policy_revision, so an agent can tell them apart.",[919,915],{"_915":916,"_917":918,"_919":920,"_921":922,"_923":924},"additionalProperties","nodes",{"_272":936,"_926":937,"_139":938},"repeat",{"_272":606,"_139":935,"_914":447},"team_id",{"_272":640,"_139":934},"activate",{"_272":931,"_139":932},"working_hours",{"_272":925,"_926":927,"_139":928},[929,930],"items",{"_272":606,"_914":447},"Named working-hours sets that defer nodes and working_hours conditions refer to by id, as an escalation path takes them.","null","array",[929,933],"Activate the new revision right away. Defaults to true.","boolean","The team's id, from list_teams.","What to do when every level has been paged and nobody answered: {times: 1..9, after: '15m'} walks the levels again.",[929,930],{"_272":606,"_914":447},"The policy's nodes in order, in the escalation path grammar, except that level targets are user targets naming current team members (membership ids from get_team) or one broadcast target ({type: broadcast}, no id) for every member. Every node needs a UUID id. See read_reference topic escalation_paths.","destructiveHint","idempotentHint","openWorldHint","readOnlyHint",{"_939":152,"_940":152,"_941":152,"_942":152},"Save a new revision of a pipeline (the entry pipeline, or one a delegate hands requests to) and activate it (unless activate is false). Blocks run in order and the first to decide wins: outcome blocks decide every request that reaches them, webhook blocks ask a service of yours (and a kind this deployment adds may ask a model), branch blocks run their own blocks only for requests their condition matches (so \"if this, then that\" is a branch holding an outcome block), and delegate blocks hand matching requests to another pipeline by key; a request that enters a branch or a delegated pipeline is decided inside it or goes to a human, and nothing after the block runs for it. A request no block decides falls through to human review, so end with an outcome block whose outcome is approve when the rest should be automatic. Every escalation path a block or the default names, and every pipeline a delegate names, must exist; activation additionally requires each to be active and refuses a delegation cycle. Call read_reference (pipeline_blocks) first.",{"_272":606,"_604":961,"_602":962,"_914":152},{"_884":71,"_885":76,"_886":71,"_887":948,"_889":76,"_890":71,"_891":892,"_882":467,"_893":73},"tyszb55mt7",{"_894":949,"_896":950,"_898":951,"_900":952,"_902":953,"_904":954},{"_488":960,"_490":71},{"_488":959,"_490":73},{"_488":958,"_490":76},{"_488":957,"_490":76},{"_488":956,"_490":71},{"_488":955,"_490":71},"For a complex revision-based mutation with no output schema, the description covers prerequisites, activation failure conditions, default-escalation inheritance, and the human-review fallback, and it routes the agent to read_reference for block specifics. Nothing an agent needs to call this correctly is missing.","Annotations only say this is a non-read-only, non-destructive, non-idempotent write. The description adds substantial behavior beyond that: activation preconditions (every named path/pipeline must exist and be active, delegation cycles refused), fall-through-to-human-review semantics, that an empty block list sends everything to human review, and that the block list wholly replaces the current one. That is exactly the extra context the annotations cannot carry.","The purpose is front-loaded in the first sentence, but the remainder is a single dense run-on paragraph that repeats block structure already covered by the schema. Much of it earns its place for a complex tool, yet it is longer than necessary and would benefit from separators given the schema and read_reference already carry detail.","Schema description coverage is 100%, so the schema already documents all six parameters in detail, and the description largely restates block/config/activate/name semantics that the schema also provides. It contributes marginal added meaning (block ordering and decision semantics), so the baseline 3 is appropriate rather than higher.","Gives clear operating context and a required prerequisite ('Call read_reference (pipeline_blocks) first'), plus a when-not path via activate=false pointing to activate_pipeline_revision. It does not explicitly contrast against create_pipeline or preview_pipeline, so it stops short of full alternative routing.","The first sentence names a specific verb+resource ('Save a new revision of a pipeline ... and activate it'), including the revision-based mutation model and the activate toggle. This cleanly separates it from create_pipeline (new pipeline) and activate_pipeline_revision (activate an existing draft), which it also cites by name.",[967,964],{"_68":963,"_964":965,"_921":966,"_967":968,"_969":970,"_971":972},{"_272":640,"_139":981},"blocks",{"_272":978,"_926":979,"_139":980},{"_272":976,"_139":977},"pipeline_key",{"_272":640,"_139":975},"default_ask_agent_owner",{"_272":933,"_139":974},"default_escalation_path",{"_272":640,"_139":973},"Escalation path key that takes human exits which name no path of their own, and requests the pipeline hands to a person by running out of blocks. Omit to inherit the defaults of the pipelines that delegate into this one, or, for the entry pipeline, to leave them to any authorized reviewer.","Send those requests to the person who owns the agent instance that made them instead, with default_escalation_path, if set, as the fallback for requests from agents without an owner (service agents); without it they go to any authorized reviewer.","The pipeline to replace ('entry' for the entry pipeline).",[929,933],"Activate the new revision right away so new requests use it. Defaults to true. False saves a draft revision you can activate later with activate_pipeline_revision.",[929,930],{"_272":606,"_914":447},"The complete ordered block list for the new revision (it replaces the current blocks, so include the ones to keep). Each block: id (key-shaped, unique across the whole pipeline including blocks inside branches), name, type (outcome | branch | delegate | webhook, or a kind this deployment adds; see read_reference), enabled: true, timeout (for example 50ms), max_attempts (1 for outcome, branch and delegate), and config for its type; a branch holds its own ordered blocks in config.blocks, a delegate names the pipeline it hands matching requests to in config.pipeline_key. See read_reference topic pipeline_blocks. An empty list sends every request to human review.","A new display name for the pipeline. Omit to keep the current one.",{"_939":152,"_940":447,"_941":152,"_942":152},"Suspend a member (signs them out everywhere and refuses their sessions, API keys and agents) or reactivate a suspended one. Refused for your own membership, the last active owner, and members your directory manages. Requires a personal API key.",{"_272":606,"_604":1001,"_602":1002,"_914":152},{"_884":71,"_885":71,"_886":71,"_887":987,"_889":76,"_890":71,"_891":988,"_882":467,"_893":73},"b0dojhy69q",{"_894":989,"_896":990,"_898":991,"_900":992,"_902":993,"_904":994},4.5,{"_488":1000,"_490":71},{"_488":999,"_490":73},{"_488":998,"_490":76},{"_488":997,"_490":71},{"_488":996,"_490":71},{"_488":995,"_490":71},"For a 3-parameter mutation with no output schema, the description covers effects, refusal conditions, and the auth prerequisite an agent needs to attempt the call correctly. Nothing material is missing.","Goes well beyond the annotations by disclosing the concrete effects of suspension (signs out everywhere, refuses sessions, API keys and agents), the refusal cases, and an auth requirement (personal API key). The reversible, idempotent nature implied matches idempotentHint=true and does not contradict destructiveHint=false.","Two tightly packed sentences, no filler. The primary action is front-loaded and the constraints and auth requirement follow in priority order.","Schema description coverage is 100%, so the schema already documents membership_id, status, and force with richer detail than the description. The description echoes the status semantics but adds nothing new for the parameters themselves; baseline 3 is appropriate.","Gives clear preconditions that select this tool's applicability (refused for your own membership, last active owner, directory-managed members) and the two legal status values. It does not explicitly route to a sibling like remove_member, so it stops short of full when-to-use/when-not guidance.","States a specific verb pair (suspend/reactivate) on a specific resource (member), and the parenthetical spells out the effect of suspension. This distinguishes it cleanly from siblings like remove_member and list_members.",[1006,742],{"_1003":1004,"_742":1005,"_1006":1007},"force",{"_272":933,"_139":1010},{"_272":640,"_139":1009},"membership_id",{"_272":640,"_139":1008},"The membership id, from list_members.","active to reactivate a suspended member, or suspended to sign them out everywhere and refuse their sessions, keys and agents until reactivated.","Suspend even while an active escalation path names the member. Without it the call is refused and lists the paths.",{"_939":152,"_940":152,"_941":152,"_942":152},"Save a new revision of a gateway server (name, URL or tunnel, and the whole scope list again) and activate it unless activate is false. Activating retires every scope the revision leaves out and drops its stored credentials; a retired scope name is refused.",{"_272":606,"_604":1029,"_602":1030,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":1016,"_889":76,"_890":73,"_891":688,"_882":467,"_893":76},"nqgume6zun",{"_894":1017,"_896":1018,"_898":1019,"_900":1020,"_902":1021,"_904":1022},{"_488":1028,"_490":73},{"_488":1027,"_490":76},{"_488":1026,"_490":76},{"_488":1025,"_490":71},{"_488":1024,"_490":73},{"_488":1023,"_490":73},"For a mutation tool with a deeply nested schema, no output schema, and a large sibling set, the description covers the decisive behavioral facts: destructive activation side effects and the non-idempotent new-revision semantics. It stops short of stating permission/auth requirements or what the call returns, leaving a small residual gap.","With annotations present, the description still adds material context the annotations do not carry: activation retires omitted scopes, drops their stored credentials, refuses retired names, and the new revision serves from the next configuration poll. The annotations correctly declare a non-read-only write, but destructiveHint=false understates the disclosed credential-dropping/retirement effect, a nuance the description fortunately corrects for the agent.","Two sentences, zero filler, and the highest-value constraint (the whole scope list must be resubmitted, and activation retires what is left out) is front-loaded before the alternative activation path is described.","Schema description coverage is 100%, so the schema already documents all six parameters, including the nested scope fields and the url/tunnel_id exclusivity. The description mainly echoes the schema (name/URL-or-tunnel/whole scope list) rather than adding syntax or format details beyond it, so the baseline of 3 applies.","It gives one real usage rule, that the entire scope list must be resubmitted on every revision, and explains the activate default. However, there is no explicit when-to-use vs when-to-use-something-else guidance against the many neighboring revision tools (create_gateway_server, activate_gateway_server_revision, archive_gateway_server). Usage is implied rather than stated.","The description names a specific verb and resource (\"Save a new revision of a gateway server\") and enumerates the content of the revision (name, URL or tunnel, whole scope list), then states the side effect (activation unless activate is false). It implicitly separates itself from activate_gateway_server_revision by bundling save+activate, but it never names a sibling or an alternative for the same effect.",[1038,68,1033],{"_92":1031,"_68":1032,"_1033":1034,"_921":1035,"_1036":1037,"_1038":1039},{"_272":640,"_139":1068},{"_272":640,"_139":1067},"scopes",{"_272":1044,"_926":1045,"_139":1046},{"_272":1042,"_139":1043},"tunnel_id",{"_272":640,"_139":1041},"server_slug",{"_272":640,"_139":1040},"The server's slug: lowercase letters, digits and dashes. Agents see its tools as __. A new slug registers the server and is taken for good, archived servers included.","The id of the tunnel the downstream server is reached through (from list_gateway_tunnels), for a server inside your own network. Give exactly one of url and tunnel_id.",[929,933],"Activate the new revision right away so the gateway serves it from its next configuration poll. Defaults to true.",[929,930],{"_272":606,"_604":1047,"_602":1048,"_914":152},"The server's scopes (named credentials), 1 to 20, in the order agents see them. The whole list again on every revision: a scope left out is retired when the revision is activated and its stored credentials are dropped.",[68,139,1052,1050],{"_68":1049,"_1050":1051,"_1052":1053,"_139":1054,"_1055":1056,"_1057":1058},{"_272":640,"_139":1066},"holder",{"_272":640,"_139":1065},"auth_mode",{"_272":640,"_139":1064},{"_272":640,"_139":1063},"oauth_scopes",{"_272":1060,"_926":1061,"_139":1062},"tunnel_entry",{"_272":640,"_139":1059},"On a server reached through a tunnel, the tunnel entry this scope calls (from list_gateway_tunnels); the tunnel holds its credential, so auth_mode is none and holder organization.",[929,930],{"_272":640},"The OAuth scopes the scope's authorization requests, for auth_mode oauth; omit to request whatever the server advertises.","What the scope's credential can do, for the agent choosing between scopes (1 to 500 characters), for example Read only access to the payments account.","none, static_headers or oauth.","organization: one credential the organization stores. member: each member connects their own on their Connections page; such a scope can be granted to personal agents only. Always organization for auth_mode none.","The scope's name: lowercase letters, digits and underscores, fixed for good. A name removed by activating a revision without it is retired and never comes back, because pipeline rules match on /request/arguments/withhuman_scope.","Display name.","The downstream server's Streamable HTTP endpoint. Give exactly one of url and tunnel_id.",{"_939":152,"_940":152,"_941":152,"_942":152},"Save a new revision of an existing escalation path (the whole document again, not a diff) and activate it unless activate is false. Pipelines that name the key pick up the active revision for new requests; requests already routed keep the revision they were pinned to.",{"_272":606,"_604":1088,"_602":1089,"_914":152},{"_884":71,"_885":73,"_886":71,"_887":1074,"_889":76,"_890":71,"_891":1075,"_882":467,"_893":73},"vjg569l6pl",{"_894":1076,"_896":1077,"_898":1078,"_900":1079,"_902":1080,"_904":1081},4.4,{"_488":1087,"_490":71},{"_488":1086,"_490":73},{"_488":1085,"_490":76},{"_488":1084,"_490":73},{"_488":1083,"_490":71},{"_488":1082,"_490":71},"For a mutation tool with full annotation coverage, 100% schema coverage, and no output schema, the description supplies the remaining behavioral gaps (revision semantics, activation default, downstream effects on pipelines and pinned requests). An agent has enough to invoke and predict consequences.","Annotations declare a non-read-only, non-idempotent, non-destructive write, and the description adds substantial context beyond that: full-document replacement rather than a diff, activation defaulting to true, that pipelines naming the key pick up the new active revision, and that already-routed requests stay pinned to their old revision. This revision-pinning semantics is exactly the kind of behavior an agent cannot infer from the schema.","Two sentences, front-loaded with the core operation and immediately followed by the activation and revision-pinning consequences. Dense but every clause carries information; no filler.","Schema description coverage is 100%, so the baseline is 3. The description reiterates the activate default ('unless activate is false'), which the schema already states, and adds no syntax or format detail beyond the schema for nodes, repeat, working_hours or path_key.","The description makes clear this applies to an existing path and bundles activation by default, which is strong context. However, it does not explicitly route the agent to activate_escalation_path_revision when activation is wanted separately, nor name create_escalation_path as the alternative for new paths.","States a specific verb and resource (save a new revision of an existing escalation path) and immediately clarifies the operation mode with '(the whole document again, not a diff)'. This distinguishes it from create_escalation_path (new path) and from activate_escalation_path_revision (activation-only sibling).",[1094,68,915],{"_68":1090,"_915":1091,"_917":1092,"_921":1093,"_1094":1095,"_923":1096},{"_272":640,"_139":1106},{"_272":1103,"_926":1104,"_139":1105},{"_272":606,"_139":935,"_914":447},{"_272":1101,"_139":1102},"path_key",{"_272":640,"_139":1100},{"_272":1097,"_926":1098,"_139":1099},[929,930],{"_272":606,"_914":447},"Named working-hours sets that defer nodes and working_hours conditions refer to by id: {id, timezone (IANA), intervals: [{weekdays: [mon, ...], start: '09:00', end: '17:30'}]}.","Stable key, lowercase letters, digits, dots, underscores or dashes (for example payments-review). It names the path from pipeline blocks and is never reused, even after archiving.",[929,933],"Activate the new revision right away. Defaults to true. A path with no active revision cannot be named by an active pipeline.",[929,930],{"_272":606,"_914":447},"The path's nodes in order: level nodes name targets and how long to wait (escalate_after) before the next level, if_else nodes branch on conditions, defer nodes wait for working hours. Every node needs a UUID id. See read_reference topic escalation_paths.","Display name people see, for example Payments review.",{"_939":152,"_940":447,"_941":152,"_942":152},"Rename one instance of an agent or switch it between active and disabled. Disabling refuses the instance's credentials until it is active again, without revoking them.",{"_272":606,"_604":1125,"_602":1126,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":1112,"_889":76,"_890":71,"_891":73,"_882":467,"_893":76},"s55y420j4c",{"_894":1113,"_896":1114,"_898":1115,"_900":1116,"_902":1117,"_904":1118},{"_488":1124,"_490":71},{"_488":1123,"_490":76},{"_488":1122,"_490":76},{"_488":1121,"_490":71},{"_488":1120,"_490":73},{"_488":1119,"_490":73},"For a 3-param, single-required, no-output-schema mutation tool whose annotations carry the safety profile, the description covers the essential behavioral twist (credentials refused, not revoked). Minor gaps: no return-value note and no permission prerequisites.","Annotations already declare mutation (readOnly=false), idempotency, and non-destructiveness. The description adds genuinely useful non-obvious behavior: disabling refuses the instance's credentials until reactivation and does so without revoking them. It doesn't cover permission requirements or timing effects.","Two sentences, front-loaded with the operations, zero filler. The credential-behavior detail is placed after the core action so an agent gets the gist immediately.","Schema coverage is 100%, so name, status, and instance_id are fully documented in the schema (including that already-made requests keep their name and that omitted fields are left unchanged). The description adds no syntax or value details beyond that, so the baseline 3 applies.","The two supported operations (rename, toggle status) are named, which implies usage, but there is no explicit when-to-use versus alternatives such as update_agent, rename_team, or create_agent_instance. No preconditions or exclusions are given.","States a specific verb and resource ('Rename one instance of an agent or switch it between active and disabled') and scopes it to a single instance, which cleanly distinguishes it from the agent-level sibling update_agent. An agent can tell what this does without opening the schema.",[1129],{"_68":1127,"_742":1128,"_1129":1130},{"_272":640,"_139":1133},{"_272":640,"_139":1132},"instance_id",{"_272":640,"_139":1131},"The instance's id, from get_agent.","active or disabled. Disabling refuses the instance's credentials until it is active again. Omit to keep the current status.","A new name, unique within the agent. Requests already made keep the name they were made with. Omit to keep the current one.",{"_939":152,"_940":447,"_941":152,"_942":152},"Rename an agent or switch it between active and disabled. Disabling refuses every instance's credentials until the agent is active again, without revoking them. The slug never changes.",{"_272":606,"_604":1153,"_602":1154,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":1139,"_889":73,"_890":71,"_891":1140,"_882":467,"_893":76},"o2s0p4k9pf",{"_894":1141,"_896":1142,"_898":1143,"_900":1144,"_902":1145,"_904":1146},4.2,{"_488":1152,"_490":71},{"_488":1151,"_490":76},{"_488":1150,"_490":73},{"_488":1149,"_490":71},{"_488":1148,"_490":73},{"_488":1147,"_490":73},"For a three-parameter mutation tool with no output schema, the description covers what changes, the immediate credential consequence, and reversibility. Return shape and name-collision behavior are left to the schema, which is acceptable, though the 'unique among live agents' constraint is never echoed in prose.","Annotations already declare idempotent, non-destructive, and not-read-only, so the safety bar is low; the description still adds real behavior: disabling refuses all instance credentials without revoking them, is reversible once active again, and leaves the slug immutable. It does not mention permission requirements or rate limits, which keeps it short of a 5.","Three short sentences with the primary action front-loaded, followed by the side effect and the invariant. No filler, no restatement of the title, nothing to cut.","Schema description coverage is 100%, so the schema already documents all three parameters and the baseline is 3. The description adds meaning beyond it by stating the slug is invariant under rename, clarifying that the effect of the name/status params is scoped and that agent_slug remains the stable identifier.","Usage is implied by the action set rather than stated: use it to rename or toggle active/disabled. It never names alternatives such as archive_agent (permanent removal) or restore_agent, so an agent must infer when this is preferred over sibling mutations.","States two specific operations (rename, switch status) on a specific resource (agent), and adds a clarifying boundary that distinguishes it from revoke_agent_credential ('without revoking them') and from slug-based identity ('the slug never changes'). An agent can tell what this tool is for without opening the schema.",[1157],{"_68":1155,"_742":1156,"_1157":1158},{"_272":640,"_139":1161},{"_272":640,"_139":1160},"agent_slug",{"_272":640,"_139":1159},"The agent's slug, from list_agents.","active or disabled. Disabling keeps every credential but refuses them all until the agent is active again. Omit to keep the current status.","A new display name, unique among live agents. Omit to keep the current one.",{"_939":152,"_940":152,"_941":447,"_942":152},"Send one signed test delivery (type webhook_endpoint.test, flagged as a preview) to a revision's URL and report what came back, waiting up to 10 seconds. It checks the receiver's signature verification too. The test is recorded in the audit log.",{"_272":606,"_604":1180,"_602":1181,"_914":152},{"_884":71,"_885":71,"_886":73,"_887":1167,"_889":76,"_890":71,"_891":1140,"_882":467,"_893":76},"u8eaz4rhaz",{"_894":1168,"_896":1169,"_898":1170,"_900":1171,"_902":1172,"_904":1173},{"_488":1179,"_490":71},{"_488":1178,"_490":76},{"_488":1177,"_490":76},{"_488":1176,"_490":71},{"_488":1175,"_490":71},{"_488":1174,"_490":73},"For a 2-parameter tool with full schema coverage and no output schema, the description covers the send, the 10s wait, and audit logging, but only vaguely says it will 'report what came back' without indicating the shape of the response (status, headers, body, signature verification result). Slightly more return detail would close the gap.","Adds substantial behavior beyond the annotations: a bounded 10-second wait, exactly one signed delivery, a preview flag, verification of the receiver's signature check, and the side effect that the test is written to the audit log. This tells the agent it is an outward-facing, non-idempotent, audit-visible action even though destructiveHint is false.","Three compact sentences, front-loaded with what is sent and where, followed by the timeout and side effects. No filler or redundancy; every clause carries information.","Schema description coverage is 100%, so both parameters (revision, endpoint_key) are already fully documented in the schema, including the omit-to-use-active-revision and draft-testing semantics. The description adds no parameter-level detail beyond that, so the baseline of 3 is appropriate.","Usage is implied by 'test delivery' and the preview flag, but the description gives no explicit when-to-use or when-not-to-use guidance, and does not name alternatives (e.g. use save_webhook_endpoint to persist, activate_..._revision to promote). The revision-vs-active choice lives in the schema, not the description.","States a specific verb (test/send), resource (webhook endpoint delivery), and scope (a single signed test delivery to a revision's URL). It is clearly distinguishable from siblings like save_webhook_endpoint, activate_webhook_endpoint_revision, and rotate_webhook_endpoint_secret.",[1184],{"_1182":1183,"_1184":1185},"revision",{"_272":1187,"_139":1188},"endpoint_key",{"_272":640,"_139":1186},"The webhook endpoint's key.","integer","The revision whose URL to test, from list_webhook_endpoint_revisions. Omit to test the active revision; a draft can be tested before it is activated.",{"_939":152,"_940":447,"_941":152,"_942":152},"Replace a team's manual member list. Requires a personal API key; organization keys cannot manage teams.",{"_272":606,"_604":1207,"_602":1208,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":1194,"_889":76,"_890":71,"_891":73,"_882":467,"_893":76},"ucfm6m7f65",{"_894":1195,"_896":1196,"_898":1197,"_900":1198,"_902":1199,"_904":1200},{"_488":1206,"_490":71},{"_488":1205,"_490":76},{"_488":1204,"_490":76},{"_488":1203,"_490":71},{"_488":1202,"_490":73},{"_488":1201,"_490":73},"For a mutation tool with full annotation coverage, a fully described schema, and no output schema, the description covers the key operational constraint an agent needs (key type). It leaves the refusal/force interaction to the schema, which documents it well, so the gap is minor.","Annotations already declare readOnlyHint=false and idempotentHint=true, so the mutation/idempotency profile is covered. The description adds non-obvious behavior beyond that: the personal-key-only auth constraint and the fact that this targets the manual list (leaving directory-synced members alone). It does not explain what the response contains or how refusals surface, but the auth caveat is real value.","Two sentences, zero padding, with the core action front-loaded and the auth caveat second. Every clause earns its place.","Schema description coverage is 100%, and the schema already explains team_id provenance (list_teams), the replace-not-merge semantics of membership_ids, and the force/escalation-policy behavior. The description adds nothing about parameters, so the baseline 3 applies.","It gives a hard prerequisite ('Requires a personal API key; organization keys cannot manage teams'), which is genuinely useful routing information, but it never names an alternative tool or the condition under which an agent should prefer update_member/remove_member over this bulk replace. Usage is implied rather than stated.","States a specific verb and resource ('Replace a team's manual member list') and qualifies it with 'manual', which separates it from directory-synced memberships and from single-member siblings like update_member/remove_member. An agent can tell at a glance that this is a bulk overwrite of the team's manually managed roster.",[919,1211],{"_1003":1209,"_919":1210,"_1211":1212},{"_272":933,"_139":1216},{"_272":640,"_139":934},"membership_ids",{"_272":1213,"_926":1214,"_139":1215},[929,930],{"_272":640},"The complete manual member list as membership ids from list_members; it replaces the current one, so include members to keep. Members a directory sync placed on the team are unaffected.","Also remove members the team's active escalation policy still names. Without it the call is refused and lists them.",{"_939":152,"_940":447,"_941":152,"_942":152},"Store the OAuth client every member signs in with on a member-held oauth scope, for an authorization server that offers no dynamic client registration. The secret is sealed and never returned. Replacing the client keeps members' tokens, though a token issued to the old client may need its member to reconnect.",{"_272":606,"_604":1235,"_602":1236,"_914":152},{"_884":71,"_885":71,"_886":71,"_887":1222,"_889":73,"_890":73,"_891":1075,"_882":467,"_893":73},"gmw04at70o",{"_894":1223,"_896":1224,"_898":1225,"_900":1226,"_902":1227,"_904":1228},{"_488":1234,"_490":73},{"_488":1233,"_490":73},{"_488":1232,"_490":73},{"_488":1231,"_490":71},{"_488":1230,"_490":71},{"_488":1229,"_490":71},"For a mutation tool with no output schema, the description covers the essential concerns an agent needs: what it stores, the DCR precondition, secret confidentiality, and the token impact of replacement. Annotations carry the safety profile, so nothing critical is missing.","Goes well beyond the annotations (readOnlyHint=false, idempotentHint=true, destructiveHint=false) by disclosing secret handling ('sealed and never returned') and the token consequence of replacement (members keep tokens, but a token issued to the old client may require reconnect). This is exactly the kind of operational context annotations cannot convey, and it is consistent with idempotentHint.","Three front-loaded sentences, each earning its place: purpose, secret-handling behavior, and replacement/migration consequence. No filler or repetition of schema facts.","Schema coverage is 100%, so the baseline is 3, but the description adds meaning about client_secret behavior ('the secret is sealed and never returned'), which clarifies the write-only nature of that parameter. The other parameters (server_slug, scope, client_id) are already well documented in the schema.","Provides a clear selecting condition: it applies 'for an authorization server that offers no dynamic client registration,' which tells the agent when this tool is the right choice. It does not explicitly name alternative tools (e.g. set_gateway_scope_headers) or state when-not-to-use, but the precondition is a strong contextual filter.","States a specific verb+resource ('Store the OAuth client ... on a member-held oauth scope') tied to a concrete use case (member sign-in). This distinguishes it from sibling tools like set_gateway_scope_headers and disconnect_gateway_scope, which act on the same scope object. The only minor gap is that the verb 'store' is slightly softer than a canonical 'set/configure'.",[1038,1237,1239],{"_1237":1238,"_1239":1240,"_1038":1241,"_1242":1243},"scope",{"_272":640,"_139":1247},"client_id",{"_272":640,"_139":1246},{"_272":640,"_139":1245},"client_secret",{"_272":640,"_139":1244},"The client secret, for a confidential client.","The server's slug, from list_gateway_servers.","The OAuth client id registered with the downstream server's authorization server.","A member-held scope whose auth_mode is oauth.",{"_939":152,"_940":447,"_941":152,"_942":152},"Store the header set an organization-held static_headers scope calls its downstream with, replacing any stored set. Values are sealed at once and never returned by any tool or page; only the gateway reads them. Ask the person for the values rather than inventing them.",{"_272":606,"_604":1266,"_602":1267,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":1253,"_889":76,"_890":73,"_891":688,"_882":467,"_893":76},"g7liwjo6hi",{"_894":1254,"_896":1255,"_898":1256,"_900":1257,"_902":1258,"_904":1259},{"_488":1265,"_490":73},{"_488":1264,"_490":76},{"_488":1263,"_490":76},{"_488":1262,"_490":71},{"_488":1261,"_490":73},{"_488":1260,"_490":73},"With no output schema and full annotation coverage, the description supplies the missing behavioral picture: irreversible-looking overwrite, secrecy, and value sourcing. It is nearly complete; only error/permission conditions are left unstated.","Annotations already declare write/idempotent/non-destructive/non-openWorld. The description adds genuinely non-obvious behavior: values are sealed immediately, are never returned by any tool or page, and are read only by the gateway. That write-only secret semantics is exactly the kind of context annotations can't convey, though it omits what happens on failure or whether the replacement is transactional.","Two sentences, zero filler, and the core mutation is front-loaded before the behavioral caveat. Every clause earns its place.","Schema coverage is 100% and each parameter is documented in the schema, so the baseline is 3. The description reinforces the replace-not-merge semantics of 'headers' and the sourcing of values, but adds no format or structural detail beyond the schema.","It gives one piece of sourcing guidance ('Ask the person for the values rather than inventing them'), which is useful. However, it does not state when to use this versus set_gateway_scope_oauth_client or disconnect_gateway_scope, nor any explicit prerequisites beyond what the schema's auth_mode constraint carries. Usage is implied rather than stated.","The description uses a specific verb+resource ('Store the header set ... replacing any stored set') and scopes it to an 'organization-held static_headers scope', which implicitly separates it from the sibling set_gateway_scope_oauth_client. It is clear what the tool does, though it never names the alternative tool explicitly.",[1038,1237,1269],{"_1237":1268,"_1269":1270,"_1038":1271},{"_272":640,"_139":1274},"headers",{"_272":606,"_139":1272,"_914":1273},{"_272":640,"_139":1245},"The complete header set the gateway sends the downstream server, for example {Authorization: 'Bearer ...'}. It replaces any stored set.",{"_272":640},"An organization-held scope whose auth_mode is static_headers.",{"_939":152,"_940":447,"_941":152,"_942":152},"Replace which of a gateway server's tools an agent may call, and with which scopes. Access decides what the agent sees and can call at all; the approval pipelines still decide every call it makes. The agent's next tool listing reflects the change, and calls check live access before they run.",{"_272":606,"_604":1294,"_602":1295,"_914":152},{"_884":73,"_885":73,"_886":73,"_887":1280,"_889":76,"_890":71,"_891":1281,"_882":467,"_893":73},"rgtpyhbkgy",{"_894":1282,"_896":1283,"_898":1284,"_900":1285,"_902":1286,"_904":1287},4.1,{"_488":1293,"_490":71},{"_488":1292,"_490":73},{"_488":1291,"_490":76},{"_488":1290,"_490":73},{"_488":1289,"_490":73},{"_488":1288,"_490":73},"With 6 fully-described params, annotations covering the safety profile, and no output schema, the definition is largely self-sufficient. The access-vs-approval distinction and propagation semantics round it out, leaving only minor gaps such as required permissions.","Annotations already declare the write/idempotent/non-destructive profile, but the description goes further by disclosing propagation semantics: 'The agent's next tool listing reflects the change, and calls check live access before they run.' That timing/liveness behavior is genuinely beyond the annotations, though permission or auth requirements are not covered.","Three tight sentences with no filler, and the core action is front-loaded ahead of the clarifying behavioral notes. Every sentence contributes, though it is slightly longer than the minimum needed.","Schema description coverage is 100%, so each parameter (mode, tools, scopes, revision) is already fully documented in the schema. The description only echoes scopes ('and with which scopes') and adds no syntax or format detail beyond it, so the baseline of 3 applies.","It draws a clear boundary between what this tool governs (access - what the agent sees and can call at all) and what it does not (the approval pipelines, which 'still decide every call it makes'), giving the agent context for choosing this over pipeline configuration. It stops short of naming an explicit alternative or a when-not-to-use condition.","The description opens with a specific verb+resource ('Replace which of a gateway server's tools an agent may call, and with which scopes'), making the write-and-replace semantics explicit and clearly distinguishable from siblings like list_agent_tool_access or list_gateway_server_tools.",[1157,1038,1296,1182],{"_1296":1297,"_1298":1299,"_1033":1300,"_1182":1301,"_1157":1302,"_1038":1303},"mode",{"_272":640,"_139":1312},"tools",{"_272":1309,"_926":1310,"_139":1311},{"_272":1306,"_926":1307,"_139":1308},{"_272":1187,"_139":1305},{"_272":640,"_139":1159},{"_272":640,"_139":1304},"The gateway server's slug, from list_gateway_servers.","The grant's current revision from list_agent_tool_access, or 0 for an agent and server never configured. A stale revision is refused.",[929,930],{"_272":640},"The scopes the agent may call with; any granted scope works with any granted tool. Empty exactly when mode is none.",[929,930],{"_272":640},"Exact downstream tool names for mode selected, as list_gateway_server_tools reports them (without the slug prefix). No wildcards.","none (no tools), selected (exactly the tools listed) or all (every current and future tool).",{"_939":152,"_940":152,"_941":152,"_942":152},"Create a webhook endpoint or save a new revision of one, and activate it unless activate is false. A new key answers with signing_secret exactly once: relay it to whoever operates the receiving service, because it is what verifies withHuman's signature and it cannot be read back. The URL must be public HTTPS.",{"_272":606,"_604":1331,"_602":1332,"_914":152},{"_884":71,"_885":73,"_886":73,"_887":1318,"_889":76,"_890":71,"_891":892,"_882":467,"_893":73},"szb2g9fvku",{"_894":1319,"_896":1320,"_898":1321,"_900":1322,"_902":1323,"_904":1324},{"_488":1330,"_490":71},{"_488":1329,"_490":73},{"_488":1328,"_490":76},{"_488":1327,"_490":73},{"_488":1326,"_490":71},{"_488":1325,"_490":73},"Given no output schema and full schema coverage, the description covers the essential return behavior (signing_secret once) and URL requirement. It omits details like what happens on failure, but for a tool with annotations and complete schema, it is nearly complete. The lack of output schema means return values are not fully specified, but the description hints at the response for new keys.","The description discloses critical behavioral traits beyond annotations: a new key returns a signing_secret exactly once and must be relayed because it cannot be read back; the URL must be public HTTPS. These are essential operational details not present in annotations (which only indicate non-read-only, non-destructive, non-idempotent). This is exactly the kind of context that prevents misuse.","The description is concise (three sentences) and front-loaded with the main action. The second sentence about signing_secret is critical but slightly parenthetical in structure. No wasted words, but could be more structured with bullet points or clearer separation of concerns. However, it remains efficient and readable.","The schema provides 100% coverage with detailed descriptions for each parameter, including 'endpoint_key' explaining the key's role in creating vs. revising. The description adds some value by clarifying the activation default and the signing_secret behavior, but most parameter semantics are already in the schema. Baseline 3 is appropriate when schema does the heavy lifting, but the description does reinforce key points.","The description says it activates unless 'activate is false', giving explicit condition for activation. However, it does not mention when to use this tool versus alternatives like activate_webhook_endpoint_revision or rotate_webhook_endpoint_secret. No explicit exclusions or alternatives are named, but the context of creating/saving revisions is clear. Sibling differentiation is implied but not spelled out.","The description clearly states a dual-purpose operation: 'Create a webhook endpoint or save a new revision of one.' It uses specific verbs ('Create', 'save') and the resource ('webhook endpoint'), and mentions activation. It distinguishes from sibling activate_webhook_endpoint_revision by noting that activation is automatic unless overridden. This is precise and identifies the tool's core function without ambiguity.",[1184,68,92],{"_92":1333,"_68":1334,"_921":1335,"_1184":1336},{"_272":640,"_139":1339},{"_272":640,"_139":1067},{"_272":1338,"_139":932},{"_272":640,"_139":1337},"Stable key a pipeline's webhook block names (config.endpoint_key). A new key creates the endpoint and mints its signing secret; an existing key gets a new revision and keeps its secret.",[929,933],"Public HTTPS URL withHuman posts the signed request to.",{"_939":447,"_940":447,"_941":152,"_942":152},"Replace a webhook endpoint's signing secret. The old secret stops working at once for every revision, so deliveries fail signature checks until the receiver has the new one. The answer carries the new secret exactly once: relay it to whoever operates the receiving service.",{"_272":606,"_604":1358,"_602":1359,"_914":152},{"_884":82,"_885":71,"_886":73,"_887":1345,"_889":76,"_890":71,"_891":474,"_882":467,"_893":73},"p7c9ysywy7",{"_894":1346,"_896":1347,"_898":1348,"_900":1349,"_902":1350,"_904":1351},{"_488":1357,"_490":71},{"_488":1356,"_490":73},{"_488":1355,"_490":76},{"_488":1354,"_490":71},{"_488":1353,"_490":82},{"_488":1352,"_490":73},"The description covers the core operation, the destructive impact on existing deliveries, and the one-time return of the secret. However, it does not clarify retry or idempotency behavior, which is especially problematic given the conflicting annotation.","The annotations declare idempotentHint=true, but the description describes a replacement operation whose 'answer carries the new secret exactly once,' implying each call generates a new secret and invalidates the previous one. This contradicts the idempotent hint and could mislead an agent about safe retry behavior.","Three sentences, front-loaded with the action, followed by the immediate consequence and the critical response behavior. Every sentence earns its place with no redundancy.","Schema coverage is 100% and the single parameter endpoint_key is fully documented in the schema. The description adds no additional meaning or format details for the parameter, so the baseline of 3 applies.","The description provides clear context for when to use the tool: when you need to replace a signing secret and can relay the new one to the receiver. It does not name alternatives or explicit exclusions, but the operational consequences make the intended usage evident.","The description states a specific verb and resource: 'Replace a webhook endpoint's signing secret.' This is unambiguous and clearly distinguishable from siblings like save_webhook_endpoint or test_webhook_endpoint. No other sibling tool performs secret rotation.",[1184],{"_1184":1360},{"_272":640,"_139":1186},{"_939":447,"_940":447,"_941":152,"_942":152},"Revoke a provisioner token for good. Instances it already created keep their own credentials.",{"_272":606,"_604":1379,"_602":1380,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":1366,"_889":76,"_890":73,"_891":474,"_882":467,"_893":79},"hf4lbwc8zl",{"_894":1367,"_896":1368,"_898":1369,"_900":1370,"_902":1371,"_904":1372},{"_488":1378,"_490":73},{"_488":1377,"_490":79},{"_488":1376,"_490":76},{"_488":1375,"_490":71},{"_488":1374,"_490":73},{"_488":1373,"_490":73},"For a one-parameter destructive tool whose annotations carry the safety profile and whose parameter is fully documented, the description supplies the key behavioral nuance (permanence, no cascade). Nothing essential is missing, though explicit usage routing would round it out.","Annotations declare destructiveHint=true and idempotentHint=true, so the safety profile is partly covered. The description adds genuinely new context: revocation is permanent ('for good') and does not cascade to already-created instances, which the annotations do not convey.","Two short sentences, no filler, with the core action and its permanence front-loaded before the non-cascade caveat.","Schema description coverage is 100% and the single parameter already documents that the id comes from get_agent. The description adds no further meaning beyond the schema, so the baseline of 3 is appropriate.","There is no explicit when-to-use or when-not-to-use guidance. The description never states the condition that should trigger a revoke versus a sibling such as revoke_agent_credential or archive_agent, leaving usage entirely to inference.","States a specific verb (Revoke) and resource (provisioner token) with an added permanence qualifier ('for good'). It is immediately clear what the tool does, though it does not name sibling alternatives like revoke_agent_credential to differentiate scope.",[1381],{"_1381":1382},"provisioner_id",{"_272":640,"_139":1383},"The provisioner token's id, from get_agent.",{"_939":447,"_940":447,"_941":152,"_942":152},"Revoke one instance credential for good: the runtime holding it can no longer make requests. To pause an instance instead, disable it with update_agent_instance.",{"_272":606,"_604":1402,"_602":1403,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":1389,"_889":76,"_890":71,"_891":1075,"_882":467,"_893":71},"at6h6c3uoj",{"_894":1390,"_896":1391,"_898":1392,"_900":1393,"_902":1394,"_904":1395},{"_488":1401,"_490":71},{"_488":1400,"_490":71},{"_488":1399,"_490":76},{"_488":1398,"_490":71},{"_488":1397,"_490":73},{"_488":1396,"_490":73},"With rich annotations, a single fully documented parameter, and no output schema, the description covers what an agent needs to invoke it safely. A minor gap remains around the instance's post-revocation state and any required permissions, but nothing blocking.","Annotations already declare destructiveHint=true and idempotentHint=true, so the safety profile is known. The description still adds real value by spelling out permanence ('for good') and the functional consequence for the runtime, which annotations cannot express. It stops short of saying what happens to the instance's other credentials or whether the revocation can be undone, so not a full 5.","Two sentences, both load-bearing: the first defines the effect, the second routes to the alternative behavior. The irreversible consequence is front-loaded and there is no filler.","Only one parameter, and the schema description is 100% covered, including the hint to obtain the id from get_agent. The description adds no syntax or format detail beyond 'one instance credential', so the baseline 3 applies.","Gives an explicit alternative and the condition that selects it: to pause rather than permanently revoke, use update_agent_instance to disable the instance. When-to-use and when-not-to-use are both covered, which is unusually complete for a single-parameter tool.","States a specific verb (Revoke), a specific resource (one instance credential), and the concrete consequence (the runtime holding it can no longer make requests). It also names the sibling tool it is not (update_agent_instance), so an agent can distinguish it from instance-management operations without opening either schema.",[1404],{"_1404":1405},"credential_id",{"_272":640,"_139":1406},"The credential's id, from get_agent.",{"_939":152,"_940":447,"_941":152,"_942":152},"Restore an archived agent to active. Credentials and provisioner tokens revoked when it was archived stay revoked; its instances can be enrolled again.",{"_272":606,"_604":1425,"_602":1426,"_914":152},{"_884":71,"_885":71,"_886":73,"_887":1412,"_889":76,"_890":71,"_891":1075,"_882":467,"_893":73},"y7dyfn61cd",{"_894":1413,"_896":1414,"_898":1415,"_900":1416,"_902":1417,"_904":1418},{"_488":1424,"_490":71},{"_488":1423,"_490":73},{"_488":1422,"_490":76},{"_488":1421,"_490":71},{"_488":1420,"_490":71},{"_488":1419,"_490":73},"For a one-parameter state-change tool with annotations covering safety and no output schema, this is nearly complete: the credential/token caveat is the key thing an agent needs. It could note whether other agent configuration is preserved, but that is a minor gap.","Beyond the annotations (idempotentHint=true, destructiveHint=false), the description discloses two non-obvious side effects: previously revoked credentials and provisioner tokens remain revoked, and instances become enrollable again. This is exactly the kind of consequence detail annotations cannot convey.","Two sentences, front-loaded with the action and followed by the consequence that matters most for a decision. No filler or restatement of the title.","Schema description coverage is 100% and the single parameter (agent_slug) is already documented as coming from list_agents. The description adds no format or generation guidance beyond the schema, so the baseline 3 applies.","The phrase 'Restore an archived agent to active' clearly communicates the precondition that the agent must currently be archived, which implicitly scopes usage. It stops short of naming archive_agent as the inverse operation or stating exclusions.","States a specific verb (restore) plus resource (agent) and the exact state transition (archived → active). An agent can distinguish this from archive_agent and update_agent without opening any schema.",[1157],{"_1157":1427},{"_272":640,"_139":1159},{"_939":152,"_940":447,"_941":152,"_942":152},"Rename a team. Escalation paths name teams by id, so nothing else changes. Refused for a team your directory manages. Requires a personal API key.",{"_272":606,"_604":1446,"_602":1447,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":1433,"_889":76,"_890":73,"_891":73,"_882":467,"_893":73},"b09o228t3m",{"_894":1434,"_896":1435,"_898":1436,"_900":1437,"_902":1438,"_904":1439},{"_488":1445,"_490":73},{"_488":1444,"_490":73},{"_488":1443,"_490":76},{"_488":1442,"_490":71},{"_488":1441,"_490":73},{"_488":1440,"_490":73},"For a two-parameter rename mutation with full schema coverage and rich annotations, the description covers purpose, side effects, restrictions, and auth. The only minor gap is that no return shape is described, though that is a low-stakes omission here.","Annotations already cover the mutation/idempotency profile (readOnlyHint=false, idempotentHint=true, destructiveHint=false). The description adds genuine context beyond them: the escalation-path reference-by-id means no cascading changes, the directory-managed refusal, and the personal-API-key auth requirement.","Four short sentences, purpose front-loaded, and each subsequent sentence adds a distinct fact (no side effects, scope refusal, auth). Nothing is redundant or padded.","Schema description coverage is 100%, so both team_id and name are already documented in the schema (including the 'unique among live teams' constraint). The description adds no parameter-level detail, so the baseline 3 applies.","Gives a clear exclusion ('Refused for a team your directory manages') and a prerequisite ('Requires a personal API key'), which is real when/when-not guidance. It stops short of naming an alternative tool for the cases it refuses.","States a specific verb and resource ('Rename a team') that an agent can act on immediately. It does not, however, differentiate itself from the sibling update_team or create_team, so the sibling-routing aspect of a 5 is missing.",[919,68],{"_68":1448,"_919":1449},{"_272":640,"_139":1450},{"_272":640,"_139":934},"The new name, unique among live teams.",{"_939":447,"_940":447,"_941":152,"_942":152},"Remove a member from the organization for good: their sessions, API keys and agent instances stop working. Refused for your own membership, the last active owner, and members your directory manages. Requires a personal API key.",{"_272":606,"_604":1469,"_602":1470,"_914":152},{"_884":71,"_885":71,"_886":71,"_887":1456,"_889":76,"_890":71,"_891":988,"_882":467,"_893":73},"efpo15537m",{"_894":1457,"_896":1458,"_898":1459,"_900":1460,"_902":1461,"_904":1462},{"_488":1468,"_490":71},{"_488":1467,"_490":73},{"_488":1466,"_490":76},{"_488":1465,"_490":71},{"_488":1464,"_490":71},{"_488":1463,"_490":71},"For a destructive mutation tool with no output schema and annotations covering the safety profile, the description supplies the destruction scope, credential requirement, and refusal cases an agent needs before calling. Nothing material is missing.","Annotations already flag destructiveHint=true and idempotentHint=true, and the description builds on that by disclosing exactly what is destroyed (sessions, API keys, agent instances), the auth requirement, and three refusal conditions. This is substantial behavioral context beyond the structured fields.","Two dense sentences with the core action and its permanence front-loaded, followed by refusal conditions and auth. No filler and nothing is buried.","Schema coverage is 100%, so membership_id (sourced from list_members) and force are already documented in the schema, including the escalation-path behavior of force. The description adds no parameter-level syntax or format beyond what the schema provides, so the baseline 3 applies.","Gives clear preconditions and exclusions: refused for your own membership, the last active owner, and directory-managed members. It also states the credential requirement ('Requires a personal API key'). It does not name an explicit alternative (e.g., update_member for reversible edits), so it stops short of a 5.","States a specific verb and resource ('Remove a member from the organization') and immediately scopes the effect as permanent, which separates it from sibling update_member. The listed downstream effects (sessions, API keys, agent instances stop working) make the action unmistakable.",[1006],{"_1003":1471,"_1006":1472},{"_272":933,"_139":1473},{"_272":640,"_139":1008},"Remove even while an active escalation path names the member. Without it the call is refused and lists the paths.",{"_939":152,"_940":152,"_941":152,"_942":447},"Read the exact grammar for pipeline blocks, conditions, or escalation path and team escalation policy documents. Read the relevant topic before writing one: unknown fields are refused and ids must be UUIDs.",{"_272":606,"_604":1492,"_602":1493,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":1479,"_889":76,"_890":73,"_891":73,"_882":467,"_893":73},"ezdrzj5b7p",{"_894":1480,"_896":1481,"_898":1482,"_900":1483,"_902":1484,"_904":1485},{"_488":1491,"_490":73},{"_488":1490,"_490":73},{"_488":1489,"_490":76},{"_488":1488,"_490":71},{"_488":1487,"_490":73},{"_488":1486,"_490":73},"For a read-only reference tool with no output schema, the description covers what is returned (the grammar), when to call it, and key validation constraints. It leaves minor gaps around the structure of the returned document, but nothing an agent needs to invoke it correctly.","Annotations already declare the safety profile (readOnlyHint=true, destructiveHint=false, openWorldHint=false). The description adds genuinely useful behavioral context beyond that: unknown fields are refused (strict validation) and ids must be UUIDs. It could still say more about the shape of the returned grammar or pagination, but the validation rules are valuable additions.","Two tightly packed sentences with no waste: the purpose and target resources come first, and the usage condition plus two validation rules follow. Nothing is redundant and the critical information is front-loaded.","Schema description coverage is 100% and the single topic parameter is already fully enumerated in the schema. The description restates the same three topic categories without adding format, default, or edge-case detail, so it does not meaningfully extend the schema.","Gives clear context for use: 'Read the relevant topic before writing one', which ties the tool to authoring workflows for pipelines, conditions and escalation paths. It stops short of naming an alternative tool or stating when this is NOT the right call.","States a specific verb (Read) and a specific resource: the exact grammar for pipeline blocks, conditions, and escalation path/team escalation policy documents. This is distinguishable from sibling instance-fetching tools like get_pipeline or get_escalation_path, though it never explicitly names those siblings to sharpen the contrast.",[1494],{"_1494":1495},"topic",{"_272":640,"_139":1496},"One of: pipeline_blocks (block envelope and each type's config), conditions (the condition tree branches, delegates and request conditions use: fields, operators, values), escalation_paths (path and team escalation policy documents: nodes, targets, timing, working hours, repeat).",{"_939":152,"_940":152,"_941":152,"_942":447},"Run a sample request through a set of blocks and report the outcome (approve, deny, human, or next), which block decided, its reason, and the escalation path a human exit would take. Each step lists a block the request reached, with the pipeline it belongs to; a branch or delegate reports enter or next for its condition, and end_of_branch when the request ran out of its blocks (for a delegate, out of the target pipeline). Delegate blocks expand to their targets' current active revisions, so the preview walks what a request would. Nothing is saved and no approval request is created. webhook blocks and other asynchronous kinds really call the endpoint or the model.",{"_272":606,"_604":1516,"_602":1517,"_914":152},{"_884":82,"_885":73,"_886":73,"_887":1502,"_889":76,"_890":71,"_891":1503,"_882":467,"_893":73},"ni6sbfgl2i",{"_894":1504,"_896":1505,"_898":1506,"_900":1507,"_902":1508,"_904":1509},3.5,{"_488":1515,"_490":71},{"_488":1514,"_490":73},{"_488":1513,"_490":76},{"_488":1512,"_490":73},{"_488":1511,"_490":82},{"_488":1510,"_490":73},"With no output schema, the description correctly explains what the report contains (outcome, deciding block, reason, escalation path, per-step block/pipeline attribution). It also covers the nested sample/blocks parameters and the async side-effect caveat; only auth/permission expectations and any rate-limit behavior are left unstated.","The description contradicts the annotations: it says webhook and other asynchronous blocks 'really call the endpoint or the model' (the schema even cites stripe__refund as a sample tool), which is external-world interaction and real side effects, while the annotations declare openWorldHint=false and readOnlyHint=true. The disclosure is otherwise rich and safety-relevant, but it directly conflicts with two declared hints, so the contradiction penalty applies.","The outcome and the dry-run guarantee are front-loaded, and the sentences are dense but each carries distinct information — the return shape has to be described in prose because there is no output schema. Some sentences (branch/delegate enter vs next vs end_of_branch) are long and could be split, but nothing is disposable.","Schema description coverage is 100%, so the schema already documents all five parameters. The description adds cross-tool semantics (blocks 'same shape update_pipeline takes', default_* 'as update_pipeline takes it') and notes that delegate blocks expand to their targets' active revisions, but the per-parameter meaning is largely schema-borne.","It establishes the dry-run context ('Nothing is saved and no approval request is created') and tells the caller to pass the current blocks from get_pipeline to test what is active, cross-referencing update_pipeline's block shape. It never names an alternative sibling to use when a real request is wanted, but the when-to-use framing is unambiguous.","States a specific verb and resource — run a sample request through a set of blocks — and enumerates the exact outcome vocabulary (approve, deny, human, next), the deciding block, its reason, and the escalation path. This clearly separates it from get_pipeline (read config), update_pipeline (write config), and test_webhook_endpoint (single-endpoint test).",[967,964,1519],{"_964":1518,"_1519":1520,"_967":1521,"_969":1522,"_971":1523},{"_272":1547,"_926":1548,"_139":1549},"sample",{"_272":606,"_604":1527,"_602":1528,"_139":1529,"_914":152},{"_272":640,"_139":1526},{"_272":933,"_139":1525},{"_272":640,"_139":1524},"The default escalation path to assume, as update_pipeline takes it.","Whether the default asks the agent's owner first, as update_pipeline takes it.","The pipeline the blocks stand in for ('entry' for the entry pipeline). Delegate blocks expand to their targets' current active revisions.",[1530],{"_1530":1531,"_1532":1533,"_1534":1535,"_1536":1537,"_1157":1538,"_1539":1540},"The request to run through the blocks. Nothing is created: no approval request, no notification.","tool",{"_272":640,"_139":1546},"context",{"_272":606,"_139":1545,"_914":447},"arguments",{"_272":606,"_139":1544,"_914":447},"agent_name",{"_272":640,"_139":1543},{"_272":640,"_139":1542},"agent_reasoning",{"_272":640,"_139":1541},"What the agent said it was trying to do, for blocks that include the agent's reasoning.","The requesting agent's slug for the sample. Defaults to 'agent'.","The requesting agent's display name for the sample. Defaults to the slug.","The tool call's arguments as the agent would send them.","Extra request context, if the adapter would send any.","The tool name the sample request asks for, for example Bash or stripe__refund.",[929,930],{"_272":606,"_914":447},"The blocks to run the sample through, in order; the same shape update_pipeline takes. Pass the current blocks from get_pipeline to test what is active.",{"_939":152,"_940":152,"_941":152,"_942":447},"List the active pipeline revisions whose webhook blocks post to this endpoint: what stands in the way of archiving it.",{"_272":606,"_604":1568,"_602":1569,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":1555,"_889":76,"_890":71,"_891":1140,"_882":467,"_893":73},"vy1x236okw",{"_894":1556,"_896":1557,"_898":1558,"_900":1559,"_902":1560,"_904":1561},{"_488":1567,"_490":71},{"_488":1566,"_490":73},{"_488":1565,"_490":76},{"_488":1564,"_490":71},{"_488":1563,"_490":73},{"_488":1562,"_490":73},"For a one-parameter read-only list tool with no output schema, the description conveys what the returned set represents (pipeline revisions that block archiving). It stops short of describing ordering or whether revisions can be returned for other endpoint states.","Annotations already declare readOnlyHint=true and destructiveHint=false, so safety is covered. The description adds a behavioral nuance beyond the annotations: only *active* pipeline revisions are returned, which tells the agent the result set is filtered rather than exhaustive.","A single sentence with the action and scope front-loaded and a colon-delimited rationale that adds value rather than padding. Nothing is wasted.","Only one parameter (endpoint_key) with 100% schema description coverage, so the schema already carries the semantics. The description refers to 'this endpoint' but adds no format or lookup detail beyond the schema.","'what stands in the way of archiving it' clearly frames the scenario that selects this tool (checking blockers before archiving a webhook endpoint). It does not name an explicit alternative or a when-not condition, so it falls short of full routing guidance.","States a specific verb and resource ('List the active pipeline revisions whose webhook blocks post to this endpoint') and even supplies the reason for the query. An agent can distinguish it from list_webhook_endpoints, list_pipeline_uses, and list_pipeline_revisions without opening any schema.",[1184],{"_1184":1570},{"_272":640,"_139":1186},{"_939":152,"_940":152,"_941":152,"_942":447},"List webhook endpoints (the URLs a pipeline's webhook block can ask for an outcome) with their key, active revision, and which pipelines use them.",{"_272":606,"_914":152},{"_884":76,"_885":71,"_886":73,"_887":1576,"_889":73,"_890":73,"_891":465,"_882":467,"_893":76},"vy0cbbhfd0",{"_894":1577,"_896":1578,"_898":1579,"_900":1580,"_902":1581,"_904":1582},{"_488":1588,"_490":73},{"_488":1587,"_490":76},{"_488":1586,"_490":73},{"_488":1585,"_490":71},{"_488":1584,"_490":76},{"_488":1583,"_490":73},"With no output schema, the description must carry the return-shape burden, and it does so by naming the key fields the agent will see. For a parameterless list tool this is nearly complete; only filtering/pagination behavior is unstated.","Annotations already declare readOnlyHint=true, destructiveHint=false and non-open-world, so the safety profile is covered. The description adds real behavioral value by disclosing the returned fields, including the cross-reference to which pipelines use each endpoint. It does not say whether archived endpoints are excluded or whether results paginate, so it adds some but not rich context.","A single front-loaded sentence with zero filler; the verb and resource come first and the parenthetical clarification earns its place by defining an ambiguous domain term. Nothing could be cut without losing meaning.","The tool takes zero parameters, so per the baseline this scores 4. The description wisely spends its words on output content instead of nonexistent inputs, describing key, active revision, and pipeline usage.","There is no explicit when-to-use or when-not-to-use guidance and no named alternative among the many webhook siblings (get_webhook_endpoint, list_webhook_endpoint_revisions, list_webhook_endpoint_uses). The plural 'List' with no parameters implies an enumeration call, but the agent must infer that this is the browse-all entry point. Minimum-viable implied usage.","States a specific verb+resource ('List webhook endpoints') and even defines the concept in a parenthetical ('the URLs a pipeline's webhook block can ask for an outcome'), which is genuinely useful domain framing. It also previews what each entry contains (key, active revision, using pipelines), which helps distinguish it from get_webhook_endpoint and list_webhook_endpoint_revisions. It stops short of naming siblings explicitly, so it stays at 4.",{"_939":152,"_940":152,"_941":152,"_942":447},"List a webhook endpoint's revisions, newest first, with which one is active. Use it to find a revision to roll back to.",{"_272":606,"_604":1607,"_602":1608,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":1594,"_889":76,"_890":71,"_891":1140,"_882":467,"_893":73},"f6clnw4eas",{"_894":1595,"_896":1596,"_898":1597,"_900":1598,"_902":1599,"_904":1600},{"_488":1606,"_490":71},{"_488":1605,"_490":73},{"_488":1604,"_490":76},{"_488":1603,"_490":71},{"_488":1602,"_490":73},{"_488":1601,"_490":73},"For a single-parameter read-only list tool with no output schema, the description covers what is returned (revisions, newest-first, active one) and why to call it. It omits pagination/result-limit behavior, which is the only notable gap.","Annotations already declare readOnlyHint=true and destructiveHint=false, so the safety profile is covered. The description adds real behavioral context beyond that: results are ordered newest-first and include which revision is currently active — return-content detail the annotations do not convey.","Two short sentences with zero filler; the scoping/ordering detail is front-loaded and the use case follows. Every clause earns its place.","Schema description coverage is 100% and the single parameter endpoint_key is documented in the schema, so the baseline is 3. The description only restates the resource the key identifies and adds no format, prefix, or lookup guidance.","'Use it to find a revision to roll back to' gives a concrete use case and implicitly routes to activate_webhook_endpoint_revision. It does not name the alternative tool or state when NOT to use this one (e.g., when you already know the revision key and should call get_webhook_endpoint_revision), so it stops short of a 5.","States a specific verb (List) and resource (a webhook endpoint's revisions), plus scope details: ordering ('newest first') and the active flag. This is clearly distinguishable from the singular sibling get_webhook_endpoint_revision and from the other list_*_revisions tools by resource.",[1184],{"_1184":1609},{"_272":640,"_139":1186},{"_939":152,"_940":152,"_941":152,"_942":447},"List every distinct tool definition across agents in one answer (name, server, schema and description, with the runtimes and instances offering it): the vocabulary to write pipeline conditions against.",{"_272":606,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":1615,"_889":73,"_890":73,"_891":1281,"_882":467,"_893":73},"hd23jvfchk",{"_894":1616,"_896":1617,"_898":1618,"_900":1619,"_902":1620,"_904":1621},{"_488":1627,"_490":73},{"_488":1626,"_490":73},{"_488":1625,"_490":73},{"_488":1624,"_490":71},{"_488":1623,"_490":73},{"_488":1622,"_490":73},"For a zero-parameter read tool with no output schema, the description covers the return contents and the dedup/single-call behavior, which is most of what an agent needs. It stops short of stating scope boundaries such as whether the listing spans all teams or only the caller's workspace, or any size limits.","Annotations already declare readOnlyHint=true, destructiveHint=false and openWorldHint=false, so the safety profile is covered. The description adds genuinely new behavior: results are deduplicated (\"distinct\") and returned in a single call, and it enumerates the payload (name, server, schema, description, runtimes, instances) in the absence of an output schema.","One sentence, front-loaded with the verb and scope, with the parenthetical payload list and the closing usage clause both earning their place. No filler or repetition of the schema.","The tool takes zero parameters, so there is no per-parameter semantics to document and the baseline of 4 applies. The description correctly implies a global, unfiltered listing rather than hinting at any input the caller might supply.","\"The vocabulary to write pipeline conditions against\" gives a clear context for when to reach for this tool, and \"in one answer\" signals it replaces multiple per-agent lookups. No exclusions or explicit alternatives are stated, so an agent must still guess how it differs from list_tool_catalog.","States a specific verb and resource (\"List every distinct tool definition across agents\") plus the scope of the aggregation, which separates it from per-agent siblings like list_agent_tool_access. It never names the closest sibling (list_tool_catalog / get_tool_catalog_entry), so the distinction from the catalog tools is left to inference.",{"_939":152,"_940":152,"_941":152,"_942":447},"List the tools agents have reported or can reach, one entry per tool and source, with its schema, the instances offering it and whether its source is available. Use it to learn the exact tool names and argument fields pipeline conditions test. The catalogue is advisory: it never changes how a request is decided.",{"_272":606,"_602":1646,"_914":152},{"_884":73,"_885":73,"_886":73,"_887":1633,"_889":76,"_890":73,"_891":465,"_882":467,"_893":76},"q9ho8wiwep",{"_894":1634,"_896":1635,"_898":1636,"_900":1637,"_902":1638,"_904":1639},{"_488":1645,"_490":73},{"_488":1644,"_490":76},{"_488":1643,"_490":76},{"_488":1642,"_490":73},{"_488":1641,"_490":73},{"_488":1640,"_490":73},"For a read-only catalogue tool with no output schema, the description explains entry granularity (one per tool and source) and entry contents well enough for an agent to interpret results. The absence of output schema is largely compensated, though pagination and total-count behavior are left to the schema.","Annotations already declare readOnlyHint=true and destructiveHint=false, so the safety profile is covered. The description adds genuine context beyond that with the advisory clause ('it never changes how a request is decided'), clarifying that the catalogue is informational and has no effect on request routing. It does not describe pagination defaults or return size beyond what the schema implies.","Three sentences that are front-loaded with the listing behavior, followed by the use case and the advisory caveat. Each sentence carries weight and there is no padding, though the middle use-case sentence slightly interrupts the flow of the two behavioral statements.","Schema description coverage is 100%, so every one of the six parameters is already documented in the schema. The description adds no filtering syntax, default values, or interaction between parameters, so the baseline 3 is appropriate.","Offers one concrete use case ('to learn the exact tool names and argument fields pipeline conditions test'), which implies when to reach for it. However, it names no alternatives, no when-not-to-use, and no prerequisites, leaving the routing decision to inference.","States a specific verb and resource ('List the tools agents have reported or can reach') and details the shape of each entry (schema, offering instances, source availability). It is clear what the tool returns, but it never distinguishes itself from close siblings like get_tool_catalog_entry or list_tool_definitions, so an agent still has to infer which is the right one.",{"_1647":1648,"_1649":1650,"_583":1651,"_742":1652,"_1653":1654,"_1157":1655},"q",{"_272":640,"_139":1661},"offset",{"_272":1187,"_139":1660},{"_272":640,"_139":1659},{"_272":640,"_139":1658},"refresh",{"_272":933,"_139":1657},{"_272":640,"_139":1656},"Only tools this agent's instances or sources offer.","Bypass the deployment's five-minute discovery cache for its own tool source. Tools a runtime reports locally refresh only through the CLI.","Only tools that are available, unavailable or stale.","Only tools from this source: builtin, mcp, or a source the deployment adds, such as gateway.","How many matching tools to skip; tools come 50 to a page.","Only tools whose name or description contains this text.",{"_939":152,"_940":152,"_941":152,"_942":447},"List teams with their id (what an escalation path's team target carries), name, member count and whether an escalation policy is active.",{"_272":606,"_602":1680,"_914":152},{"_884":76,"_885":71,"_886":73,"_887":1667,"_889":76,"_890":71,"_891":688,"_882":467,"_893":76},"hosyzv1psw",{"_894":1668,"_896":1669,"_898":1670,"_900":1671,"_902":1672,"_904":1673},{"_488":1679,"_490":71},{"_488":1678,"_490":76},{"_488":1677,"_490":76},{"_488":1676,"_490":71},{"_488":1675,"_490":76},{"_488":1674,"_490":73},"Given the tool's simplicity, full schema coverage, and annotation-based safety profile, the description is complete enough. It does not mention the 'status' filter or pagination, but those are either covered by the schema or not required for this simple listing operation.","Annotations already declare readOnlyHint, destructiveHint=false, and openWorldHint=false, so the safety profile is covered. The description adds value by specifying returned fields and the id's semantic role, but omits any behavioral details like pagination, rate limits, or default filtering behavior.","A single, front-loaded sentence that packs useful information without any filler. Every clause contributes to understanding the tool's output.","Schema description coverage is 100%, so the single 'status' parameter is fully documented in the schema. The description adds no extra meaning about the parameter, so the baseline score of 3 is appropriate.","The description implies usage through 'List teams' but provides no explicit when-to-use guidance or mention of alternatives. An agent can infer it should be used to enumerate teams, but no exclusions or conditions are given.","States a specific verb ('List') and resource ('teams') and enumerates the returned fields, including the meaning of the team id in an escalation path context. This distinguishes it from sibling tools like get_team and the various other list_* tools without ambiguity.",{"_742":1681},{"_272":640,"_139":1682},"Which to list: live (the default), archived, or all.",{"_939":152,"_940":152,"_941":152,"_942":447},"List the permission policies a team holds, which every member inherits live: each policy's permissions, scope and revision. Requires a personal API key.",{"_272":606,"_604":1701,"_602":1702,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":1688,"_889":76,"_890":73,"_891":688,"_882":467,"_893":76},"fqt144ywmc",{"_894":1689,"_896":1690,"_898":1691,"_900":1692,"_902":1693,"_904":1694},{"_488":1700,"_490":73},{"_488":1699,"_490":76},{"_488":1698,"_490":76},{"_488":1697,"_490":71},{"_488":1696,"_490":73},{"_488":1695,"_490":73},"With no output schema, the description compensates by naming the returned fields (permissions, scope, revision) and the live-inheritance behavior. For a one-parameter read tool this is nearly complete; only pagination/result-size expectations are unstated.","Annotations already establish a safe read (readOnlyHint=true, destructiveHint=false), yet the description adds meaningful context beyond them: the policies are inherited 'live' by every member, the response shape is enumerated, and an auth requirement (personal API key) is disclosed. This is genuinely additive, though it doesn't cover pagination or ordering.","One sentence, front-loaded with the action and resource, followed by return contents and the auth prerequisite. No redundant or filler text.","With a single parameter at 100% schema description coverage (the schema states the team_id is 'from list_teams'), the schema already carries the parameter semantics. The description adds nothing further about the id, so the baseline 3 applies.","Gives one concrete prerequisite, 'Requires a personal API key', which is real usage guidance. However, it never says when to reach for this tool versus list_member_permission_policies, list_permissions, or get_team, so the selection context is only implied.","Names a specific verb and resource (list team permission policies) and specifies the returned fields (permissions, scope, revision). It implicitly distinguishes itself from the sibling list_member_permission_policies by scoping to policies 'a team holds', but never names that sibling, so differentiation relies on inference.",[919],{"_919":1703},{"_272":640,"_139":934},{"_939":152,"_940":152,"_941":152,"_942":447},"List a team's escalation policy revisions, newest first, with which one is active.",{"_272":606,"_604":1722,"_602":1723,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":1709,"_889":76,"_890":73,"_891":688,"_882":467,"_893":76},"fwqwrw0cy1",{"_894":1710,"_896":1711,"_898":1712,"_900":1713,"_902":1714,"_904":1715},{"_488":1721,"_490":73},{"_488":1720,"_490":76},{"_488":1719,"_490":76},{"_488":1718,"_490":71},{"_488":1717,"_490":73},{"_488":1716,"_490":73},"With no output schema, the description usefully discloses sort order and the active-revision marker, which is what an agent most needs to interpret results. It is silent on pagination or result-size limits, a minor gap for a single-parameter list tool.","Annotations already declare readOnlyHint=true and destructiveHint=false, so the safety profile is covered. The description adds genuinely new behavioral context: results are ordered newest first and the active revision is flagged, which tells the agent how to interpret the response without an output schema.","One sentence, zero filler, with the ordering and active-revision detail front-loaded alongside the core action. Nothing could be removed without losing information.","Schema description coverage is 100% and the single team_id parameter is already documented in the schema ('The team's id, from list_teams.'). The description only restates the team scoping, adding no syntax or format detail beyond the schema, so baseline 3 applies.","Usage is implied by the verb and resource: fetch the revision history for one team's escalation policy. There is no explicit guidance on when to use this versus get_team_escalation_policy_revision or activate_team_escalation_policy_revision, and no stated prerequisites beyond the team_id.","States a specific verb (List) and fully-qualified resource (a team's escalation policy revisions), which cleanly separates it from siblings like list_escalation_path_revisions and list_gateway_server_revisions. It does not explicitly route against sibling tools, but the resource naming is unambiguous.",[919],{"_919":1724},{"_272":640,"_139":934},{"_939":152,"_940":152,"_941":152,"_942":447},"List the predefined roles (owner, admin, reviewer, auditor) with the permissions each grants. Assigning roles and permission policies is done by a person in the app. Requires a personal API key.",{"_272":606,"_914":152},{"_884":73,"_885":71,"_886":71,"_887":1730,"_889":73,"_890":73,"_891":1140,"_882":467,"_893":73},"cfqdxmqpcm",{"_894":1731,"_896":1732,"_898":1733,"_900":1734,"_902":1735,"_904":1736},{"_488":1742,"_490":73},{"_488":1741,"_490":73},{"_488":1740,"_490":73},{"_488":1739,"_490":71},{"_488":1738,"_490":73},{"_488":1737,"_490":71},"For a zero-parameter, read-only list of a fixed set of roles with no output schema, the description is sufficient: it names the returned entities and the permission detail each carries, plus the auth prerequisite. Nothing an agent needs to call it correctly is absent.","Annotations already cover the safety profile (readOnlyHint=true, destructiveHint=false, openWorldHint=false), so the bar is lower. The description adds real context beyond them: the auth requirement ('Requires a personal API key') and the fact that role assignment is out of band. It stops short of describing return format or the fixed nature of the list.","Three short sentences, each earning its place: purpose/return content first, scope boundary second, auth requirement last. No filler and the core capability is front-loaded.","The tool takes zero parameters, so there is nothing for the description to disambiguate; baseline 4 applies. No parameter-level meaning is needed or missing.","The sentence 'Assigning roles and permission policies is done by a person in the app' clearly bounds the tool's scope (read-only enumeration, no mutation path via the API), which is a genuine when-not-to-use signal. It does not route the agent to any alternative tool, so it falls short of a 5.","Specific verb ('List') plus resource ('predefined roles') and even enumerates the exact roles (owner, admin, reviewer, auditor), which separates it from the sibling list_permissions that lists granular permission definitions. It does not explicitly name the sibling it differs from, so it stops short of a 5.",{"_939":152,"_940":152,"_941":152,"_942":447},"List the delegate blocks in active pipeline revisions that hand requests to this pipeline: what stands in the way of archiving it.",{"_272":606,"_604":1761,"_602":1762,"_914":152},{"_884":76,"_885":71,"_886":73,"_887":1748,"_889":76,"_890":73,"_891":688,"_882":467,"_893":73},"ok6k5ada1j",{"_894":1749,"_896":1750,"_898":1751,"_900":1752,"_902":1753,"_904":1754},{"_488":1760,"_490":73},{"_488":1759,"_490":73},{"_488":1758,"_490":76},{"_488":1757,"_490":71},{"_488":1756,"_490":76},{"_488":1755,"_490":73},"For a single-parameter read tool with no output schema and full annotation coverage, the description tells the agent what will be returned conceptually (delegate blocks pointing at this pipeline) and why it matters. The remaining gap is that it doesn't hint at the shape or size of the result set, but nothing essential is missing.","Annotations already declare readOnlyHint=true, destructiveHint=false and openWorldHint=false, so the safety profile is covered. The description's only behavioral addition is the scope constraint that only ACTIVE revisions are considered (drafts excluded) and the framing that results represent archive blockers. That is useful but thin, and it does not explain why idempotentHint is false for a list operation.","One sentence, front-loaded with the verb and resource, with the motivational clause placed last. No filler or redundancy.","Schema description coverage is 100% and the schema itself already explains pipeline_key, including the special 'entry' value. The description adds no syntax or format detail about the parameter beyond restating that it identifies the pipeline receiving delegated requests, so baseline 3 applies.","The 'what stands in the way of archiving it' clause supplies a clear use context: consult this before archiving a pipeline. It does not state exclusions or name an alternative tool for other reference queries, so it stops short of a 5.","The description gives a specific verb (List) and a precisely qualified resource: delegate blocks in active pipeline revisions that hand requests to this pipeline. It also states the resource's meaning (archive blockers), which separates it from siblings like list_pipeline_revisions or get_pipeline_revision_references, though it does not name those alternatives directly.",[967],{"_967":1763},{"_272":640,"_139":1764},"The pipeline's key. 'entry' is the entry pipeline every request starts in; any other key is a pipeline a delegate block hands requests to.",{"_939":152,"_940":152,"_941":152,"_942":447},"List every approval pipeline: the entry pipeline every request starts in, then every pipeline a delegate block can hand requests to, each with its key, name, active revision number, the delegate blocks that hand requests to it, and history.",{"_272":606,"_602":1783,"_914":152},{"_884":73,"_885":73,"_886":71,"_887":1770,"_889":76,"_890":73,"_891":474,"_882":467,"_893":79},"ekka8d40t7",{"_894":1771,"_896":1772,"_898":1773,"_900":1774,"_902":1775,"_904":1776},{"_488":1782,"_490":73},{"_488":1781,"_490":79},{"_488":1780,"_490":76},{"_488":1779,"_490":73},{"_488":1778,"_490":73},{"_488":1777,"_490":71},"For a read-only list tool with one fully-documented optional parameter and no output schema, the description supplies the domain model and the return shape an agent needs to interpret results. Nothing essential is missing.","Annotations already declare readOnlyHint=true and destructiveHint=false, so the safety profile is covered by structured fields. The description adds genuine value beyond that by enumerating the returned content — key, name, active revision number, delegating blocks, and history — which matters because there is no output schema. It still omits auth requirements and pagination behavior.","One front-loaded sentence with no filler; the enumeration of returned fields is dense but each item earns its place because no output schema exists to carry them. Slightly long, but efficient.","Schema coverage is 100% and the single status parameter is fully documented in the schema (live default, archived, all). The description never mentions status, and its phrase \"List every approval pipeline\" is slightly at odds with the default live-only behavior dictated by the schema. Baseline 3 is appropriate since the schema carries the semantics.","There is no when-to-use guidance and no named alternative. Nothing tells the agent when to call this versus get_pipeline (single lookup), list_pipeline_revisions, or list_pipeline_uses. The only usage signal is the implicit purpose of a list operation.","The description states a clear verb and resource (\"List every approval pipeline\") and goes further by explaining the domain structure: an entry pipeline plus every pipeline reachable via delegate blocks. That is far more specific than a tautology. It does not, however, explicitly separate itself from siblings like get_pipeline or list_pipeline_revisions, so peer differentiation is left implicit.",{"_742":1784},{"_272":640,"_139":1682},{"_939":152,"_940":152,"_941":152,"_942":447},"List a pipeline's revisions, newest first, with which one is active. Use it to find a revision to roll back to.",{"_272":606,"_604":1803,"_602":1804,"_914":152},{"_884":76,"_885":71,"_886":73,"_887":1790,"_889":76,"_890":71,"_891":73,"_882":467,"_893":73},"hgwx4ag36u",{"_894":1791,"_896":1792,"_898":1793,"_900":1794,"_902":1795,"_904":1796},{"_488":1802,"_490":71},{"_488":1801,"_490":73},{"_488":1800,"_490":76},{"_488":1799,"_490":71},{"_488":1798,"_490":76},{"_488":1797,"_490":73},"With a single schema-documented parameter and no output schema, an agent has enough to call it, and the description usefully sketches the return shape. It could still say more about volume or limits for pipelines with many revisions, which is the only real gap.","Annotations already cover the safety profile (readOnlyHint=true, destructiveHint=false, openWorldHint=false), so the burden is lighter. The description adds genuine behavioral context beyond them by disclosing result ordering and the active-revision marker, but says nothing about result size, pagination, or behavior on an unknown pipeline_key.","Two tight sentences with zero filler; the what and the ordering constraint are front-loaded, and the usage hint follows immediately.","Only one parameter and schema description coverage is 100%; the schema itself explains that 'entry' is the entry pipeline and other keys are delegate-block targets. The description adds no parameter meaning beyond that, so the baseline 3 applies.","\"Use it to find a revision to roll back to\" gives a concrete decision context that implicitly routes the agent toward activate_pipeline_revision as the follow-up. It stops short of naming alternatives or stating exclusions (e.g., use get_pipeline_revision when you already know the revision id), so it is clear but not exhaustive.","Specific verb (list) plus resource (a pipeline's revisions) with two scope qualifiers: ordering (newest first) and content (which revision is active). This cleanly separates it from get_pipeline_revision, activate_pipeline_revision, and list_pipeline_uses in the sibling set.",[967],{"_967":1805},{"_272":640,"_139":1764},{"_939":152,"_940":152,"_941":152,"_942":447},"List every permission key this deployment knows, with what it allows, who can hold it, the kind of resource it applies to and whether it is dangerous. Use it to tell a person which grant to make, for example pipeline.write at pipeline: for a team that owns a delegated pipeline. Requires a personal API key that can read permission policies.",{"_272":606,"_914":152},{"_884":73,"_885":73,"_886":73,"_887":1811,"_889":73,"_890":73,"_891":73,"_882":467,"_893":73},"owny3dl0bk",{"_894":1812,"_896":1813,"_898":1814,"_900":1815,"_902":1816,"_904":1817},{"_488":1823,"_490":73},{"_488":1822,"_490":73},{"_488":1821,"_490":73},{"_488":1820,"_490":73},{"_488":1819,"_490":73},{"_488":1818,"_490":73},"With no parameters, no output schema, and read-only annotations, the description carries the remaining burden well by describing the fields each permission entry exposes and the credential required. It stops short of covering pagination or ordering behavior, which would matter for a full permission catalog.","Annotations already declare readOnlyHint=true and destructiveHint=false, so safety is covered, and the description usefully adds an authorization prerequisite the annotations do not carry ('requires a personal API key that can read permission policies'). The idempotentHint=false is odd for a listing operation, but the description does not contradict it, and it omits any note on pagination or result size.","Two sentences, front-loaded with what the tool returns before the usage and auth notes. Every clause adds information, though the return-field enumeration and the example make the first sentence fairly dense.","The tool takes zero parameters, which is the baseline-4 case; there is nothing for the description to disambiguate. The schema itself has no properties, so no semantics are missing or misrepresented.","It states a concrete use case - mapping a person to the right grant, with a worked example ('pipeline.write at pipeline: for a team that owns a delegated pipeline'). That is clear context for when to reach for it, but it names no alternatives or exclusions, so an agent cannot tell when to prefer it over the policy/role listing tools.","It names a specific verb and resource ('List every permission key this deployment knows') and enumerates what each entry conveys, so the agent knows exactly what this returns. It does not, however, explicitly distinguish itself from nearby siblings such as list_roles, list_member_permission_policies, or list_team_permission_policies, so sibling differentiation is left to inference.",{"_939":152,"_940":152,"_941":152,"_942":447},"List the organization's members with the membership id (the id an escalation path's user target carries), display name, email, status, roles and team ids.",{"_272":606,"_914":152},{"_884":76,"_885":71,"_886":73,"_887":1829,"_889":73,"_890":71,"_891":688,"_882":467,"_893":79},"gh7apt2tc1",{"_894":1830,"_896":1831,"_898":1832,"_900":1833,"_902":1834,"_904":1835},{"_488":1841,"_490":71},{"_488":1840,"_490":79},{"_488":1839,"_490":73},{"_488":1838,"_490":71},{"_488":1837,"_490":76},{"_488":1836,"_490":73},"With no output schema, the description usefully enumerates the returned fields (membership id, display name, email, status, roles, team ids), which is exactly what an agent needs. It falls short only on unstated result-size/pagination behavior for a parameterless list call.","Annotations already declare readOnlyHint=true and destructiveHint=false, so the safety profile is covered. The description adds valuable context by enumerating returned fields and cross-referencing the membership id used by escalation path user targets, but says nothing about result size, pagination, or whether the full org membership is returned with zero input parameters.","A single tight sentence, front-loaded with the verb+resource and then the returned attributes. Nothing is wasted and nothing is buried.","The tool takes zero parameters, so the baseline is 4. The description correctly adds no parameter information, and instead spends its words on the return payload, which is the right allocation.","There is no guidance on when to use this tool versus related tools (list_teams, list_roles, list_permissions, list_member_permission_policies). Usage is only implied by the verb 'List'. No exclusions or prerequisites are given.","States a specific verb and resource ('List the organization's members') and further pins down the entity by listing the fields returned. This is clearly distinct from siblings like list_teams, list_roles, and list_permissions; an agent can select it without ambiguity.",{"_939":152,"_940":152,"_941":152,"_942":447},"List the permission policies granted directly to one member: each policy's permissions, the scope they apply at and its revision. Grants the member inherits from teams are on list_team_permission_policies. Requires a personal API key.",{"_272":606,"_604":1860,"_602":1861,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":1847,"_889":76,"_890":71,"_891":1075,"_882":467,"_893":71},"dc6d51yv4s",{"_894":1848,"_896":1849,"_898":1850,"_900":1851,"_902":1852,"_904":1853},{"_488":1859,"_490":71},{"_488":1858,"_490":71},{"_488":1857,"_490":76},{"_488":1856,"_490":71},{"_488":1855,"_490":73},{"_488":1854,"_490":73},"For a one-parameter read-only list tool with full annotation coverage, this is nearly complete: scope, return shape, sibling routing and auth are all present. Missing pagination/empty-result behavior and the fact that there is no output schema, so the return contract is only described obliquely.","Annotations already declare readOnlyHint=true, destructiveHint=false and openWorldHint=false, so the safety profile is covered. The description adds genuinely new behavior: the personal-API-key auth requirement and the restriction to direct (non-inherited) grants. It does not mention pagination or result caps, which is the only remaining gap for a list endpoint.","Three short sentences, front-loaded with the verb and scope, then the sibling routing, then the auth caveat. Every clause carries information an agent needs to choose or invoke the tool; nothing is redundant.","Schema coverage is 100% and the single parameter already documents itself ('The membership id, from list_members'), so the baseline is 3. The description's 'granted directly to one member' reinforces that membership_id is a single-member scope, but adds no format or syntax detail beyond the schema.","Explicitly names the alternative ('Grants the member inherits from teams are on list_team_permission_policies') and the condition that selects it (direct vs inherited). It also adds a prerequisite the agent cannot infer elsewhere: 'Requires a personal API key.' Both when-to-use and a hard gating condition are present.","States a specific verb and resource ('List the permission policies granted directly to one member') and even enumerates what each entry contains (permissions, scope, revision). The word 'directly' carves out a precise scope that separates it from inherited grants, so an agent can identify it without opening the schema.",[1006],{"_1006":1862},{"_272":640,"_139":1008},{"_939":152,"_940":152,"_941":152,"_942":447},"List the tunnels the organization runs in its own network, with their live connections and the entries they serve: the tunnel_id and tunnel_entry values a server reached through a tunnel names.",{"_272":606,"_602":1881,"_914":152},{"_884":76,"_885":73,"_886":73,"_887":1868,"_889":76,"_890":73,"_891":471,"_882":462,"_893":79},"m2pge1y1k1",{"_894":1869,"_896":1870,"_898":1871,"_900":1872,"_902":1873,"_904":1874},{"_488":1880,"_490":73},{"_488":1879,"_490":79},{"_488":1878,"_490":76},{"_488":1877,"_490":73},{"_488":1876,"_490":76},{"_488":1875,"_490":73},"With no output schema, the description usefully sketches the return shape (tunnel_id, tunnel_entry, live connections, entries served). For a one-optional-param read tool with full annotation coverage, this is close to complete, missing only pagination/volume expectations.","Annotations already declare readOnlyHint=true and destructiveHint=false, so the safety profile is covered. The description adds that results include live connections and the entries served, which is useful content context, but says nothing about pagination, permissions, or auth requirements.","A single front-loaded sentence that leads with the primary purpose. The trailing clause ('the tunnel_id and tunnel_entry values a server reached through a tunnel names') is grammatically tangled and slightly hard to parse, costing some clarity.","Schema coverage is 100% and the single status parameter is fully documented in the schema. The description only glancingly reinforces the live default via 'their live connections' and adds no syntax or filtering detail beyond the schema, so baseline 3 applies.","No when-to-use or when-not-to-use guidance is given. The status parameter implies a live/archived/all distinction, but the description never states when an agent should reach for this tool versus list_gateway_servers or get_gateway_server.","States a specific verb (List) and resource (tunnels the organization runs in its own network), and explains what a tunnel carries (live connections, entries served). This clearly separates it from the gateway-server siblings, though it never names a sibling directly.",{"_742":1882},{"_272":640,"_139":1682},{"_939":152,"_940":152,"_941":152,"_942":447},"List the tools a gateway server's downstream advertises, by their own names, as the gateway last listed them: the union, and per scope the tools that scope's credential lists, or why it could not. A scope marked refreshing is being listed again in the background (never listed, changed, or an hour old): call again shortly for the new listing. Use the names for set_agent_tool_access and pipeline conditions. Requires a personal API key.",{"_272":606,"_604":1901,"_602":1902,"_914":152},{"_884":71,"_885":73,"_886":73,"_887":1888,"_889":76,"_890":73,"_891":1281,"_882":467,"_893":73},"fajr1u2q0e",{"_894":1889,"_896":1890,"_898":1891,"_900":1892,"_902":1893,"_904":1894},{"_488":1900,"_490":73},{"_488":1899,"_490":73},{"_488":1898,"_490":76},{"_488":1897,"_490":73},{"_488":1896,"_490":71},{"_488":1895,"_490":73},"With no output schema, the description does the work of describing returns (the union plus per-scope lists and failure reasons), and it covers auth and refresh behavior. It is nearly complete; only the exact shape of the per-scope failure payload is left implicit.","Despite annotations already marking it read-only, the description discloses what the annotations cannot: the result is a cached snapshot ('as the gateway last listed them'), the background refresh semantics for scopes marked refreshing with a retry hint, that failures are reported per scope with reasons, and that a personal API key is required. This is substantive behavioral context beyond the structured fields.","Every sentence carries information and the listing/refresh/retry points are ordered logically, but the opening sentence is densely packed with colon-separated clauses that take a second read. No filler, only slight structural overhead.","Schema coverage is 100% and the single server_slug parameter is already well documented in the schema, including its origin (list_gateway_servers) and the __ prefix. The description adds no parameter-level detail, so the baseline 3 applies.","Gives a concrete downstream use ('Use the names for set_agent_tool_access and pipeline conditions') and a retry instruction for scopes marked refreshing. It stops short of stating when NOT to use it or naming a sibling to prefer instead, so it is clear context without explicit exclusions.","States a specific verb and resource ('list the tools a gateway server's downstream advertises, by their own names') and pins the scope to a single gateway server, which separates it from catalog-wide siblings like list_tool_catalog. It never names an alternative tool, so an agent still has to infer the routing on its own.",[1038],{"_1038":1903},{"_272":640,"_139":1904},"The server's slug, from list_gateway_servers. It prefixes the server's tools as __ for agents.",{"_939":152,"_940":152,"_941":152,"_942":447},"List the MCP servers the organization registered on the MCP Gateway: slug, name, URL or tunnel, active revision, each scope (named credential) with whether its credential is stored and any warnings about its OAuth grant. Secret values are never returned.",{"_272":606,"_602":1924,"_914":152},{"_884":73,"_885":73,"_886":73,"_887":1910,"_889":76,"_890":71,"_891":1911,"_882":467,"_893":76},"ditng5lbo4",{"_894":1912,"_896":1913,"_898":1914,"_900":1915,"_902":1916,"_904":1917},3.9,{"_488":1923,"_490":71},{"_488":1922,"_490":76},{"_488":1921,"_490":76},{"_488":1920,"_490":73},{"_488":1919,"_490":73},{"_488":1918,"_490":73},"With no output schema, the description correctly compensates by enumerating the returned fields and the secret-handling guarantee. It is nearly complete for a read-only list; only pagination/volume behavior is unaddressed.","Annotations already declare readOnlyHint=true and destructiveHint=false, so the safety profile is covered. The description adds genuinely useful context beyond that: secret values are never returned, and each scope surfaces credential-storage state and OAuth-grant warnings. No pagination or ordering behavior is disclosed.","One front-loaded sentence that names the action and resource before enumerating returned fields. Dense but every clause carries information; the field enumeration is justified because no output schema exists.","Schema description coverage is 100% and the sole 'status' parameter is fully documented in the schema (live default, archived, all). The description does not mention the filter at all, so it adds nothing beyond the schema — the baseline 3 applies.","The listing intent is implied by the verb and the org-level scope, but the description never says when to reach for this versus get_gateway_server (single server) or the revision/tool listing siblings, nor does it surface the status filter as a usage consideration. Adequate but with a clear routing gap.","States a specific verb (List) and resource (MCP servers registered on the MCP Gateway) plus the exact fields returned (slug, name, URL/tunnel, active revision, scopes, credential status, OAuth warnings). This is clearly distinguishable from get_gateway_server, list_gateway_server_revisions, and list_gateway_server_tools by scope.",{"_742":1925},{"_272":640,"_139":1682},{"_939":152,"_940":152,"_941":152,"_942":447},"List a gateway server's revisions, newest first, with which one is active. Use it to find a revision to roll back to.",{"_272":606,"_604":1944,"_602":1945,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":1931,"_889":76,"_890":73,"_891":73,"_882":467,"_893":73},"n0rmuxqr5h",{"_894":1932,"_896":1933,"_898":1934,"_900":1935,"_902":1936,"_904":1937},{"_488":1943,"_490":73},{"_488":1942,"_490":73},{"_488":1941,"_490":76},{"_488":1940,"_490":71},{"_488":1939,"_490":73},{"_488":1938,"_490":73},"With no output schema, the description must carry return semantics, and it does supply the ordering and active-flag detail; pagination is handled by the cursor parameter in the schema. Nothing critical is missing for a read-only listing tool, though a note on pagination behavior would close the gap.","Annotations already declare readOnlyHint=true and destructiveHint=false, so the safety profile is covered. The description still adds value beyond them by disclosing the return ordering (newest first) and the presence of an active-revision indicator, which is the only place an agent learns the response shape since there is no output schema.","Two short sentences, front-loaded with what the tool returns and followed by the use case. No redundant phrasing or filler.","Schema description coverage is 100% and both parameters (server_slug, cursor) are documented in the schema. The description adds no parameter-level meaning beyond that, so the baseline of 3 applies.","'Use it to find a revision to roll back to' gives a clear primary scenario, implying the prerequisite that you need this before calling activate_gateway_server_revision. It stops short of explicit when-not guidance or naming the alternative tools for fetching revision details.","States a specific verb and resource ('List a gateway server's revisions') plus two concrete output facts: newest-first ordering and which revision is active. That is far more than a restatement of the name, though it does not name the sibling get_gateway_server_revision to sharpen the boundary.",[1038],{"_1946":1947,"_1038":1948},"cursor",{"_272":640,"_139":1949},{"_272":640,"_139":1245},"The next_cursor of the previous page.",{"_939":152,"_940":152,"_941":152,"_942":447},"List the active pipeline revisions that name this path, as a block's path or as the default: what stands in the way of archiving it.",{"_272":606,"_604":1968,"_602":1969,"_914":152},{"_884":73,"_885":71,"_886":71,"_887":1955,"_889":76,"_890":73,"_891":1911,"_882":467,"_893":76},"o2ql3gcorc",{"_894":1956,"_896":1957,"_898":1958,"_900":1959,"_902":1960,"_904":1961},{"_488":1967,"_490":73},{"_488":1966,"_490":76},{"_488":1965,"_490":76},{"_488":1964,"_490":71},{"_488":1963,"_490":73},{"_488":1962,"_490":71},"There is no output schema, so the description must convey what comes back, and it does: active pipeline revisions, categorized by how they reference the path. Combined with the stated archival motivation and the annotation-covered safety profile, an agent has everything needed to call it correctly.","Annotations already cover the safety profile (readOnlyHint=true, destructiveHint=false), so the bar is lower. The description adds real behavioral context beyond them: results are limited to ACTIVE pipeline revisions, and two distinct reference kinds are enumerated (block's path vs. default). It does not address the odd idempotentHint=false on a read-only lister, nor pagination.","One sentence, verb first, no filler. The colon construction packs the return semantics and the motivation into the same breath without redundancy.","There is a single required parameter with 100% schema description coverage ('The escalation path's key.'), so the schema carries the burden and baseline 3 applies. The description says only 'this path' and adds no format, key syntax, or lookup guidance beyond the schema.","The trailing clause 'what stands in the way of archiving it' implies the use case (pre-flight check before archive_escalation_path), but usage is only implied, never stated as an instruction. No when-not guidance and no explicit alternative is named.","The description states a specific verb and resource: list the active pipeline revisions that name this escalation path. It also clarifies the naming modes (a block's path or the default), which distinguishes it from generic listers such as list_escalation_paths or list_pipeline_revisions. It never names a sibling directly, so differentiation is inferred rather than explicit.",[1094],{"_1094":1970},{"_272":640,"_139":1971},"The escalation path's key.",{"_939":152,"_940":152,"_941":152,"_942":447},"List escalation paths with their key, name, active revision and latest revision.",{"_272":606,"_602":1990,"_914":152},{"_884":76,"_885":73,"_886":76,"_887":1977,"_889":76,"_890":73,"_891":471,"_882":462,"_893":79},"ade5o8o0m2",{"_894":1978,"_896":1979,"_898":1980,"_900":1981,"_902":1982,"_904":1983},{"_488":1989,"_490":73},{"_488":1988,"_490":79},{"_488":1987,"_490":76},{"_488":1986,"_490":73},{"_488":1985,"_490":76},{"_488":1984,"_490":76},"For a simple read-only list tool with a fully documented single parameter, this is close to adequate. The main gaps are the absence of sibling disambiguation and any note on ordering/pagination, which matter in a catalog dense with similar list_* tools.","Annotations already declare readOnlyHint=true and destructiveHint=false, so the safety profile is covered. The description adds the returned field set (key, name, active/latest revision), which is mildly useful context, but it says nothing about pagination, ordering, or result size.","A single tight sentence with the action front-loaded and no filler. Efficient, though the space freed by brevity is not spent on any differentiating or usage context.","Schema coverage is 100%, so the lone 'status' parameter (live/archived/all) is fully documented in the schema. The description does not extend or clarify parameter behavior, so the baseline 3 applies.","There is no when-to-use guidance and no named alternative. With siblings such as list_escalation_path_revisions and list_escalation_path_uses in the catalog, the description should say which of these to pick, but it offers nothing.","Starts with a specific verb+resource ('List escalation paths') and names the fields returned, so the purpose is unambiguous. It does not, however, distinguish itself from close siblings like get_escalation_path, list_escalation_path_revisions, or list_escalation_path_uses, leaving the agent to infer the difference.",{"_742":1991},{"_272":640,"_139":1682},{"_939":152,"_940":152,"_941":152,"_942":447},"List an escalation path's revisions, newest first, with which one is active.",{"_272":606,"_604":2010,"_602":2011,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":1997,"_889":76,"_890":71,"_891":73,"_882":467,"_893":76},"qko8t6yl4m",{"_894":1998,"_896":1999,"_898":2000,"_900":2001,"_902":2002,"_904":2003},{"_488":2009,"_490":71},{"_488":2008,"_490":76},{"_488":2007,"_490":76},{"_488":2006,"_490":71},{"_488":2005,"_490":73},{"_488":2004,"_490":73},"For a simple read-only list tool with one fully documented parameter and annotations covering safety, the description is largely complete. It explains ordering and active status but does not mention pagination or result limits.","The annotations already declare readOnlyHint=true and destructiveHint=false. The description adds useful behavioral context beyond that by stating results are newest first and indicating which revision is active, though it does not cover pagination or permission requirements.","The description is a single front-loaded sentence with no filler. Every clause adds useful information: what is listed, ordering, and active-revision marking.","Schema description coverage is 100%, and the sole path_key parameter is already documented in the schema as the escalation path's key. The description adds no further parameter semantics beyond what the schema provides, so this is the baseline for high schema coverage.","The purpose implies when to use it, but there is no explicit guidance about alternatives such as get_escalation_path_revision or when not to use this list tool. Usage is inferable from the name and description rather than stated.","The description states a specific verb and resource: list an escalation path's revisions. It also adds ordering and active-status scope, making it easy to distinguish from get_escalation_path_revision and from list_* tools for other resource types.",[1094],{"_1094":2012},{"_272":640,"_139":1971},{"_939":152,"_940":152,"_941":152,"_942":447},"List the event types this organization's audit log holds, the vocabulary for list_audit_events' event_types filter.",{"_272":606,"_914":152},{"_884":76,"_885":71,"_886":73,"_887":2018,"_889":73,"_890":71,"_891":1140,"_882":467,"_893":73},"ykxkzb0sox",{"_894":2019,"_896":2020,"_898":2021,"_900":2022,"_902":2023,"_904":2024},{"_488":2030,"_490":71},{"_488":2029,"_490":73},{"_488":2028,"_490":73},{"_488":2027,"_490":71},{"_488":2026,"_490":76},{"_488":2025,"_490":73},"There is no output schema, but 'vocabulary' strongly implies a list of event-type identifiers, which is sufficient for a zero-argument enumeration tool. A brief note on return shape or whether values are stable would make it fully complete.","Annotations already declare readOnlyHint=true, openWorldHint=false, and destructiveHint=false, so the safety profile is covered. The description adds the org-scoping detail ('this organization's audit log'), but says nothing about the volume or shape of the returned list. Adequate given the annotation coverage.","A single sentence, front-loaded with the action and resource, with the sibling relationship appended as a compact clause. No waste.","The tool takes zero parameters and the schema is a bare empty object, so there is no parameter semantics burden to carry. Baseline for a 0-param tool applies.","The description frames this as the vocabulary lookup for list_audit_events' event_types filter, which implies the correct usage pattern (call this first to discover valid filter values). It stops short of explicit when-not guidance, but the alternative is clearly signposted.","States a specific verb and resource ('List the event types this organization's audit log holds') and explicitly distinguishes itself from list_audit_events by naming the sibling and its filter. An agent can tell exactly which of the two audit tools to call.",{"_939":152,"_940":152,"_941":152,"_942":447},"Search the organization's audit log, newest first by default: who did what to which subject and how it ended, every approval decision, pipeline outcome, configuration change and API key action included. Filters combine; page with next_cursor. total_count is how many events match.",{"_272":606,"_602":2049,"_914":152},{"_884":73,"_885":73,"_886":71,"_887":2036,"_889":73,"_890":73,"_891":1911,"_882":467,"_893":76},"zlfzknf3yr",{"_894":2037,"_896":2038,"_898":2039,"_900":2040,"_902":2041,"_904":2042},{"_488":2048,"_490":73},{"_488":2047,"_490":76},{"_488":2046,"_490":73},{"_488":2045,"_490":73},{"_488":2044,"_490":73},{"_488":2043,"_490":71},"For a 13-parameter, read-only search with no output schema, the description covers the essentials: what is searchable, filter semantics, ordering default, and the paging and count fields an agent will see in the response. Combined with 100% schema coverage, an agent has what it needs to call this correctly.","Annotations already establish the safe read profile (readOnlyHint=true, destructiveHint=false, openWorldHint=false). The description adds behavior beyond that: default newest-first ordering, that filters are conjunctive, the cursor-based paging mechanism, and the meaning of total_count. It doesn't cover rate limits or result caps beyond limit's 500 bound in the schema.","Three dense sentences, front-loaded with the purpose before the mechanics. Nearly every clause carries information, though 'who did what to which subject and how it ended' is somewhat rhetorical filler versus the sharper coverage list that follows.","Schema description coverage is 100%, so the baseline is 3. The description earns above baseline by clarifying cross-parameter semantics that the schema does not: that multiple filters combine (AND) rather than override, that default sort is newest-first, and that total_count reflects the full match set rather than the page.","It gives operational guidance for using the tool ('Filters combine; page with next_cursor') but no when-to-use-this-vs-alternatives guidance. There is no mention of when to reach for get_audit_event on a single event, or of calling list_audit_event_types to discover valid values as implied by the event_types param description.","States a specific verb and resource ('Search the organization's audit log') and enumerates the event coverage (approval decisions, pipeline outcomes, configuration changes, API key actions), which makes the scope concrete. It does not explicitly name siblings like get_audit_event or list_audit_event_types, so the agent must infer the list-vs-single distinction from naming alone.",{"_1647":2050,"_2051":2052,"_2053":2054,"_2055":2056,"_2057":2058,"_1946":2059,"_2060":2061,"_2062":2063,"_2064":2065,"_2066":2067,"_2068":2069,"_2070":2071,"_2072":2073},{"_272":640,"_139":2088},"sort",{"_272":640,"_139":2087},"limit",{"_272":1187,"_139":2086},"order",{"_272":640,"_139":2085},"to_at",{"_272":640,"_139":2084},{"_272":640,"_139":2083},"from_at",{"_272":640,"_139":2082},"outcome",{"_272":640,"_139":2081},"actor_id",{"_272":640,"_139":2080},"actor_type",{"_272":640,"_139":2079},"subject_id",{"_272":640,"_139":2078},"event_types",{"_272":2075,"_926":2076,"_139":2077},"subject_type",{"_272":640,"_139":2074},"Only events about this kind of thing, such as approval_request, membership or request_approval_pipeline.",[929,930],{"_272":640},"Only events of these types, from list_audit_event_types.","Only events about one subject; requires subject_type. An agent is given by its slug, everything else by its id.","Only events by this kind of actor: human, agent, system or api_key.","Only events by one actor: a membership or user id, an agent's slug, or an API key's id.","Only events whose action ended this way: approved, denied, expired, cancelled, human or blocked.","Only events at or after this RFC 3339 time.","The next_cursor of the previous page; keep the same filters, sort and order.","Only events before this RFC 3339 time.","desc for newest first (the default) or asc.","How many events per page, 1 to 500. Defaults to 100.","occurred (the default), event_type or actor.","Free text over the event type, the actor's name and id, the subject id, the event id and the event's data.",{"_939":152,"_940":152,"_941":152,"_942":447},"List an agent's grants on gateway servers: per server, the mode (none, selected or all), the tools and scopes granted, and the revision set_agent_tool_access needs. Every instance of the agent inherits them.",{"_272":606,"_604":2107,"_602":2108,"_914":152},{"_884":73,"_885":73,"_886":73,"_887":2094,"_889":76,"_890":71,"_891":1281,"_882":467,"_893":73},"f6vt0aivtd",{"_894":2095,"_896":2096,"_898":2097,"_900":2098,"_902":2099,"_904":2100},{"_488":2106,"_490":71},{"_488":2105,"_490":73},{"_488":2104,"_490":76},{"_488":2103,"_490":73},{"_488":2102,"_490":73},{"_488":2101,"_490":73},"For a low-complexity, single-parameter read with no output schema, the description adequately characterizes what is returned (per-server mode, tools, scopes, revision) and the inheritance behavior. It's sufficient to call and interpret the tool, with only minor gaps in pagination/format detail.","Annotations already establish readOnlyHint=true and destructiveHint=false, so safety is covered. The description adds genuine beyond-annotation context: the concrete return shape (per-server mode of none/selected/all, tools, scopes, revision) and the inheritance rule that 'every instance of the agent inherits them'.","A single dense sentence front-loads the action and resource, then enumerates the payload fields and the inheritance fact. Nothing is wasted, though the nested enumeration makes it slightly heavy to parse.","Schema coverage is 100% and the single agent_slug parameter is fully documented in the schema, including its source (list_agents). The description adds no additional parameter-level meaning, so the baseline 3 applies for a one-param tool whose schema does the work.","The description cross-references the sibling set_agent_tool_access by noting it returns 'the revision set_agent_tool_access needs', which tells the agent this is the read-before-write step for updating access. However, it stops short of an explicit when-to-use/when-not statement or named alternatives.","States a specific verb (List) and resource (an agent's grants on gateway servers), and clarifies the granularity (per server, with mode/tools/scopes/revision). This clearly differentiates it from siblings like list_gateway_servers and list_gateway_server_tools, which operate at the server level rather than the agent-grant level.",[1157],{"_1157":2109},{"_272":640,"_139":1159},{"_939":152,"_940":152,"_941":152,"_942":447},"List the organization's agents: slug (what pipelines and conditions refer to), name, kind, status, and recent request counts.",{"_272":606,"_602":2128,"_914":152},{"_884":76,"_885":71,"_886":73,"_887":2115,"_889":76,"_890":73,"_891":474,"_882":467,"_893":76},"m376v8fwnv",{"_894":2116,"_896":2117,"_898":2118,"_900":2119,"_902":2120,"_904":2121},{"_488":2127,"_490":73},{"_488":2126,"_490":76},{"_488":2125,"_490":76},{"_488":2124,"_490":71},{"_488":2123,"_490":76},{"_488":2122,"_490":73},"For a zero-required-parameter list tool with full schema coverage, the description is close to sufficient, and it usefully substitutes for the absent output schema by naming the returned fields. It stops short of covering pagination or org-wide permission constraints.","Annotations already declare readOnlyHint=true and destructiveHint=false, so the safety profile is covered. The description adds the returned attribute set, which is genuinely useful given there is no output schema, but it says nothing about pagination, result caps, or permission scoping for the org-wide listing.","A single front-loaded sentence with no filler; the parenthetical clarifying what a slug is for earns its place by adding domain meaning.","Only one parameter, and schema description coverage is 100% — the status enum values (live/archived/all) and the default are fully documented in the schema. The description adds no syntax or format detail beyond that, so baseline 3 applies.","Usage is only implied: the parenthetical about slug hints at how agents are referenced elsewhere, and the schema's status parameter hints at the archive workflow. There is no explicit statement of when to use this instead of get_agent, list_members, or the archive/restore tools.","States a specific verb and resource (\"List the organization's agents\") and enumerates the attributes returned, so the agent knows this is a collection read rather than a single-agent fetch. It does not explicitly contrast itself with the singular get_agent sibling, which keeps it out of the top band.",{"_742":2129},{"_272":640,"_139":1682},{"_939":152,"_940":152,"_941":152,"_942":447},"Read one webhook endpoint revision, active or not.",{"_272":606,"_604":2148,"_602":2149,"_914":152},{"_884":76,"_885":71,"_886":73,"_887":2135,"_889":76,"_890":73,"_891":474,"_882":467,"_893":76},"e9ccjd4tla",{"_894":2136,"_896":2137,"_898":2138,"_900":2139,"_902":2140,"_904":2141},{"_488":2147,"_490":73},{"_488":2146,"_490":76},{"_488":2145,"_490":76},{"_488":2144,"_490":71},{"_488":2143,"_490":76},{"_488":2142,"_490":73},"For a simple, fully-schema-covered read tool with annotations carrying the safety profile and no output schema, the description is nearly sufficient. A small gap remains around what a 'revision' snapshot contains and how it differs from the live endpoint.","Annotations already declare readOnlyHint=true and destructiveHint=false, so the safety profile is covered. The description adds the useful fact that inactive/historical revisions are readable ('active or not'), but does not explain the non-idempotent hint or anything about the returned payload.","A single short sentence with zero waste, front-loading the verb, resource, and scope qualifier.","Schema description coverage is 100%, so both endpoint_key and revision are documented in the schema, including the pointer to list_webhook_endpoint_revisions. The description adds no further parameter meaning, so the baseline 3 applies.","There is no explicit when-to-use or when-not-to-use guidance. The schema's note that 'revision' comes 'from list_webhook_endpoint_revisions' implies a list-then-read workflow, but the description itself gives only implied routing between this and the list/get siblings.","The description states a specific verb ('Read') and resource ('one webhook endpoint revision'), and the phrase 'active or not' clarifies that revision history is in scope. It distinguishes itself from get_webhook_endpoint and list_webhook_endpoint_revisions implicitly, though it never names a sibling explicitly.",[1184,1182],{"_1182":2150,"_1184":2151},{"_272":1187,"_139":2152},{"_272":640,"_139":1186},"The revision number, from list_webhook_endpoint_revisions.",{"_939":152,"_940":152,"_941":152,"_942":447},"Read a webhook endpoint's active revision: name and URL. The signing secret is never returned here; it is shown once when the endpoint is created or its secret rotated.",{"_272":606,"_604":2171,"_602":2172,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":2158,"_889":76,"_890":73,"_891":688,"_882":467,"_893":76},"wse2ffbik5",{"_894":2159,"_896":2160,"_898":2161,"_900":2162,"_902":2163,"_904":2164},{"_488":2170,"_490":73},{"_488":2169,"_490":76},{"_488":2168,"_490":76},{"_488":2167,"_490":71},{"_488":2166,"_490":73},{"_488":2165,"_490":73},"There is no output schema, so the description must describe the return surface, and it does name the returned fields while explicitly excluding the secret. It stops short of covering error conditions or what 'active revision' means if no revision is active.","Annotations already establish that this is a safe, non-destructive read, so the bar is lower. The description still adds real behavioral value by ruling out the signing secret ('never returned here') and explaining it is exposed only at creation or rotation, which steers the agent away from a futile fetch.","Two tight sentences with the scope of the read front-loaded and the secret caveat immediately after. No filler or redundancy.","With a single parameter at 100% schema description coverage, the schema already carries the semantics of endpoint_key. The description adds nothing about the key's format or provenance, so the baseline 3 applies.","Usage is only implied: the tool reads the current/active configuration of an endpoint. There is no statement of when to prefer it over get_webhook_endpoint_revision, list_webhook_endpoints, or save_webhook_endpoint, and no prerequisites or exclusions are given.","States a specific verb and resource ('Read a webhook endpoint's active revision') and even enumerates the returned fields (name and URL), which is sharper than a bare 'get'. It does not explicitly differentiate itself from close siblings like get_webhook_endpoint_revision or list_webhook_endpoints, so it stops short of a 5.",[1184],{"_1184":2173},{"_272":640,"_139":1186},{"_939":152,"_940":152,"_941":152,"_942":447},"Read one tool catalogue entry with its full input schema and description.",{"_272":606,"_604":2193,"_602":2194,"_914":152},{"_884":76,"_885":71,"_886":73,"_887":2179,"_889":76,"_890":73,"_891":2180,"_882":462,"_893":79},"aml68gx2hj",{"_894":2181,"_896":2182,"_898":2183,"_900":2184,"_902":2185,"_904":2186},3.4,{"_488":2192,"_490":73},{"_488":2191,"_490":79},{"_488":2190,"_490":76},{"_488":2189,"_490":71},{"_488":2188,"_490":76},{"_488":2187,"_490":73},"For a low-complexity read with no output schema, the description tells the agent what comes back (full input schema plus description) and annotations cover safety, so it is nearly complete. The remaining gap is the agent_slug viewpoint behavior, which is only discoverable from the schema.","Annotations already declare readOnlyHint=true, destructiveHint=false and openWorldHint=false, so the safety profile is covered. The description adds the one genuinely new behavioral fact — that the response contains the entry's full input schema and description — which is useful given there is no output schema. It stops short of noting that agent_slug changes the result, so it is only partially additive.","A single tight sentence with the resource and the returned payload front-loaded. Every clause earns its place and there is no filler.","Schema description coverage is 100%, so both parameters are already documented, including that the id comes from list_tool_catalog and that agent_slug scopes the read to a particular agent's view. The description adds no syntax, format, or scoping detail beyond that, which is the expected baseline when the schema carries the load.","There is no when-to-use guidance and no named alternative, even though list_tool_catalog is an obvious sibling and the id parameter's schema text points back to it. The description says nothing about prerequisites or when this tool is preferable to reading the catalogue via a list call.","The description states a specific verb ('Read') and a specific resource ('one tool catalogue entry'), with the scope limiter 'one' implicitly contrasting with the plural list_tool_catalog sibling. It never names that sibling explicitly, so an agent must infer the list-vs-get routing from the wording alone.",[20],{"_20":2195,"_1157":2196},{"_272":640,"_139":2198},{"_272":640,"_139":2197},"Read the entry as this agent sees it.","The catalogue entry's id, from list_tool_catalog.",{"_939":152,"_940":152,"_941":152,"_942":447},"Read one team escalation policy revision, active or not, with warnings for targets who have since left the team.",{"_272":606,"_604":2217,"_602":2218,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":2204,"_889":76,"_890":71,"_891":73,"_882":467,"_893":76},"nhl1ycgl0k",{"_894":2205,"_896":2206,"_898":2207,"_900":2208,"_902":2209,"_904":2210},{"_488":2216,"_490":71},{"_488":2215,"_490":76},{"_488":2214,"_490":76},{"_488":2213,"_490":71},{"_488":2212,"_490":73},{"_488":2211,"_490":73},"For a simple two-parameter read tool with complete schema coverage and annotations that cover safety, the description is nearly sufficient. It could improve by briefly indicating the return shape or warning format, but it already supplies the main non-obvious behavioral caveat.","Annotations already establish that this is a non-destructive, read-only, non-open-world operation. The description adds useful context beyond annotations: revisions may be active or not, and the response can include warnings about targets who have since left the team.","The description is a single well-structured sentence with no wasted wording. The key behavioral detail about warnings is front-loaded alongside the primary read purpose.","Both parameters are fully documented in the input schema, including where team_id and revision come from. The description does not add parameter-level meaning beyond the schema, which is acceptable given 100% schema description coverage.","The description implies usage by naming the operation as reading a single revision, and the schema points to list_team_escalation_policy_revisions as the source for the revision number. However, it gives no explicit when-to-use guidance, exclusions, or direct comparison to get_team_escalation_policy or the list tool.","The description states a specific verb and resource: read one team escalation policy revision. It explicitly covers both active and inactive revisions and adds the distinctive warning behavior, which separates it from list and activation siblings without needing their schemas.",[919,1182],{"_919":2219,"_1182":2220},{"_272":640,"_139":934},{"_272":1187,"_139":2221},"The policy revision number, from list_team_escalation_policy_revisions.",{"_939":152,"_940":152,"_941":152,"_942":447},"Read a team's active escalation policy: how a level targeting the team pages its members, with warnings for targets who have left. Not found when no policy is active; such a level then reaches every member at once.",{"_272":606,"_604":2240,"_602":2241,"_914":152},{"_884":73,"_885":73,"_886":73,"_887":2227,"_889":76,"_890":73,"_891":465,"_882":467,"_893":76},"sid4vffxd0",{"_894":2228,"_896":2229,"_898":2230,"_900":2231,"_902":2232,"_904":2233},{"_488":2239,"_490":73},{"_488":2238,"_490":76},{"_488":2237,"_490":76},{"_488":2236,"_490":73},{"_488":2235,"_490":73},{"_488":2234,"_490":73},"With no output schema, the description carries the burden of explaining what is returned, and it does sketch the returned policy's meaning plus the missing-policy case. For a single-parameter read tool this is close to complete, though the actual shape of the returned policy is only described narratively.","Annotations already declare readOnlyHint=true and destructiveHint=false, so the safety profile is covered. The description adds value beyond that by disclosing the not-found case (\"Not found when no policy is active\") and the fallback semantics that apply in that case, which an agent needs to interpret an empty result correctly.","Two tight sentences with the core purpose front-loaded and edge-case behavior second. The final clause (\"such a level then reaches every member at once\") is compressed to the point of mild ambiguity, which keeps it just short of ideal.","There is only one parameter and schema description coverage is 100%; the schema already documents team_id including its source (list_teams). The description adds no additional syntax or format meaning, so the baseline of 3 applies.","Usage is implied by the attribute \"active\" rather than stated: an agent can infer this is the current-state read versus the revision reader. There is no explicit when-to-use/when-not guidance or routing to sibling tools such as get_team or list_team_escalation_policy_revisions.","States a specific verb and resource (\"Read a team's active escalation policy\") and goes further by explaining what the resource actually means: how a level targeting the team pages its members. The word \"active\" implicitly separates it from get_team_escalation_policy_revision, though that sibling is never named explicitly.",[919],{"_919":2242},{"_272":640,"_139":934},{"_939":152,"_940":152,"_941":152,"_942":447},"Read one team and its members.",{"_272":606,"_604":2261,"_602":2262,"_914":152},{"_884":76,"_885":73,"_886":73,"_887":2248,"_889":76,"_890":73,"_891":1503,"_882":467,"_893":76},"m3f5nmpobo",{"_894":2249,"_896":2250,"_898":2251,"_900":2252,"_902":2253,"_904":2254},{"_488":2260,"_490":73},{"_488":2259,"_490":76},{"_488":2258,"_490":76},{"_488":2257,"_490":73},{"_488":2256,"_490":76},{"_488":2255,"_490":73},"With one fully documented parameter, full annotation coverage, and no output schema, the description covers the essentials and even notes the members payload is included. Minor gaps remain around what else the team object contains and how many members are returned.","Annotations already declare readOnlyHint=true, destructiveHint=false, idempotentHint=false, and openWorldHint=false, so the safety profile is covered structurally. The description adds only one piece of real behavioral context — that the member roster is included in the return — with no detail on member pagination or team fields.","A single front-loaded sentence with zero filler. It is efficient, though terse enough that it leans toward under-specification for a tool with a composite return shape.","Schema description coverage is 100%: team_id is documented, including where to obtain it ('from list_teams'). The description adds nothing beyond the schema, so the baseline 3 applies.","There is no explicit when-to-use guidance or named alternative (e.g. list_teams for bulk, get_team_escalation_policy for policy data). The singular scope implies the lookup-one-team-by-id case, but the agent must infer this.","States a specific verb (read) and resource (team) with singular scope ('one team'), which implicitly distinguishes it from list_teams. It does not explicitly name or route against any sibling, so it stops short of a 5.",[919],{"_919":2263},{"_272":640,"_139":934},{"_939":152,"_940":152,"_941":152,"_942":447},"Tell one approval request's story from its audit events: the pipeline's conclusion and the block that asked for a person, every AI judge assessment, the escalation steps and the claim history. Needs only request.read on the request, not audit.read.",{"_272":606,"_604":2282,"_602":2283,"_914":152},{"_884":73,"_885":73,"_886":73,"_887":2269,"_889":76,"_890":71,"_891":1281,"_882":467,"_893":73},"ibpihn7nio",{"_894":2270,"_896":2271,"_898":2272,"_900":2273,"_902":2274,"_904":2275},{"_488":2281,"_490":71},{"_488":2280,"_490":73},{"_488":2279,"_490":76},{"_488":2278,"_490":73},{"_488":2277,"_490":73},{"_488":2276,"_490":73},"With no output schema, the description carries the burden of describing returns, and it does list the timeline's components plus the auth requirement. Ordering, pagination, and error behavior are unaddressed, leaving a small gap.","Annotations already declare readOnlyHint=true and destructiveHint=false, so safety is covered. The description adds genuinely new behavioral context: an unexpected permission requirement (request.read rather than audit.read) and a preview of the composite content returned. It does not cover pagination or ordering, which keeps it from a 5.","One dense sentence that front-loads the core action and closes with the permission detail. Slight flourish in 'tell one approval request's story,' but no wasted sentences.","Schema description coverage is 100% and the single request_id parameter is already documented in the schema with a concrete example (an audit event's subject_id). The description adds no additional parameter meaning, so the baseline of 3 is appropriate.","Explicitly states the authorization context: 'Needs only request.read on the request, not audit.read,' which implicitly tells an agent it can use this instead of audit-scoped tooling. It stops short of naming a specific alternative tool or a when-not-to-use condition, so it is clear but not fully routing.","States a specific verb and resource — assembling one approval request's timeline from its audit events — and enumerates what the timeline contains (pipeline conclusion, human-approval block, AI judge assessments, escalation steps, claim history). This clearly distinguishes it from the raw event retrieval siblings like get_audit_event and list_audit_events.",[2284],{"_2284":2285},"request_id",{"_272":640,"_139":2286},"The approval request's id, for example an audit event's subject_id when subject_type is approval_request.",{"_939":152,"_940":152,"_941":152,"_942":447},"List everything a pipeline revision names (escalation paths, webhook endpoints, delegated pipelines, an edition's block dependencies) with each one's state: what stands in the way of activating it.",{"_272":606,"_604":2305,"_602":2306,"_914":152},{"_884":76,"_885":73,"_886":73,"_887":2292,"_889":76,"_890":73,"_891":1503,"_882":467,"_893":76},"uazofpnjee",{"_894":2293,"_896":2294,"_898":2295,"_900":2296,"_902":2297,"_904":2298},{"_488":2304,"_490":73},{"_488":2303,"_490":76},{"_488":2302,"_490":76},{"_488":2301,"_490":73},{"_488":2300,"_490":76},{"_488":2299,"_490":73},"With no output schema, the description usefully sketches what is returned (references plus their state and activation blockers), and the two required params are fully covered by the schema. Only pagination/size behavior and explicit sibling routing are missing.","Annotations already declare readOnlyHint=true and destructiveHint=false, so the safety profile is covered. The description does add useful return-content context (each reference's state and activation blockers), which matters since there is no output schema, but it says nothing about volume, pagination, or how blockers are represented.","One dense sentence with the resource and the enumerated reference types front-loaded, and the value of the returned state ('what stands in the way of activating it') at the end. No filler, though the parenthetical enumeration makes it heavy.","Schema description coverage is 100% and both parameters (pipeline_key, revision) are documented in the schema, including the 'entry' convention and the pointer to list_pipeline_revisions. The description adds no parameter-level detail beyond what the schema already carries, so the baseline 3 applies.","Usage is only implied: the mention of 'what stands in the way of activating it' suggests this is a pre-activation check against activate_pipeline_revision, but the description never states when to call it versus get_pipeline_revision or list_pipeline_uses. No exclusions or prerequisites are given.","The description states a specific verb+resource ('List everything a pipeline revision names') and enumerates the concrete reference kinds it covers (escalation paths, webhook endpoints, delegated pipelines, block dependencies). It distinguishes itself implicitly from the bare get_pipeline_revision / list_pipeline_revisions siblings, though it never names a sibling outright.",[967,1182],{"_1182":2307,"_967":2308},{"_272":1187,"_139":2310},{"_272":640,"_139":2309},"The pipeline's key ('entry' for the entry pipeline).","The revision number, from list_pipeline_revisions.",{"_939":152,"_940":152,"_941":152,"_942":447},"Read one pipeline revision, active or not, with its blocks.",{"_272":606,"_604":2329,"_602":2330,"_914":152},{"_884":76,"_885":71,"_886":73,"_887":2316,"_889":76,"_890":73,"_891":2180,"_882":462,"_893":79},"p5ripe2wp1",{"_894":2317,"_896":2318,"_898":2319,"_900":2320,"_902":2321,"_904":2322},{"_488":2328,"_490":73},{"_488":2327,"_490":79},{"_488":2326,"_490":76},{"_488":2325,"_490":71},{"_488":2324,"_490":76},{"_488":2323,"_490":73},"For a simple two-parameter read tool with full schema coverage and complete safety annotations, the definition covers purpose, scope, and return content adequately. With no output schema, a slightly richer note about what 'blocks' contains would be ideal, but nothing essential to invoking it correctly is missing.","Annotations already declare this a safe, non-destructive, non-open-world read, so the safety burden is lifted. The description usefully adds that inactive revisions are retrievable and that blocks are returned, but says nothing about idempotency (annotations set idempotentHint=false, which is not explained) or about auth/permission behavior.","A single front-loaded sentence with no filler. The scope qualifier and return content are packed into one clause and nothing is wasted.","Schema coverage is 100% and both parameters carry descriptions, including the workflow hint that revision numbers come from list_pipeline_revisions. The description adds no parameter-level meaning, so the baseline 3 applies.","The description gives no when-to-use guidance, no prerequisites, and names no alternative. 'Active or not' hints at scope but does not tell the agent when to reach for this versus get_pipeline or list_pipeline_revisions.","States a specific verb and resource ('Read one pipeline revision') and scopes it ('active or not'), which implicitly distinguishes a single-item read from the list_pipeline_revisions sibling. It does not explicitly name how it differs from get_pipeline or get_pipeline_revision_references, so it stops short of a 5.",[967,1182],{"_1182":2331,"_967":2332},{"_272":1187,"_139":2310},{"_272":640,"_139":2309},{"_939":152,"_940":152,"_941":152,"_942":447},"Read a pipeline's active revision with its blocks in order and its default escalation path. Blocks come back in the shape update_pipeline accepts, so you can edit and resubmit them.",{"_272":606,"_604":2351,"_602":2352,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":2338,"_889":76,"_890":73,"_891":688,"_882":467,"_893":76},"jdzwfk34nd",{"_894":2339,"_896":2340,"_898":2341,"_900":2342,"_902":2343,"_904":2344},{"_488":2350,"_490":73},{"_488":2349,"_490":76},{"_488":2348,"_490":76},{"_488":2347,"_490":71},{"_488":2346,"_490":73},{"_488":2345,"_490":73},"With no output schema, the description usefully characterizes the return (ordered blocks, default escalation path, update_pipeline-compatible shape). It stops short of describing failure modes or auth requirements, which is a modest gap for a single-parameter read tool.","Annotations already declare readOnlyHint=true and destructiveHint=false, so safety is covered. The description adds genuinely non-obvious behavior: the returned blocks are in the exact shape update_pipeline accepts, making the round-trip safe. It still omits error behavior for an unknown key and any permission requirements.","Two sentences, no filler, and the most important scoping fact ('active revision') is front-loaded. The second sentence adds the round-trip contract that an agent needs for the common follow-up action.","Schema coverage is 100% and the schema's own description of pipeline_key already explains the 'entry' semantics versus delegate targets, which the tool description does not repeat or extend. Baseline 3 is appropriate since the schema carries the parameter meaning.","Usage is implied by the workflow note that blocks can be edited and resubmitted via update_pipeline, which effectively tells the agent this is the read half of an edit loop. However, it never states when to prefer this over get_pipeline_revision, list_pipeline_revisions, or preview_pipeline, and offers no exclusions.","States a specific verb and resource ('Read a pipeline's active revision') plus exactly what is returned: ordered blocks and the default escalation path. The word 'active' implicitly separates it from get_pipeline_revision (a specific revision) and list_pipelines, but no sibling is named, so an agent must infer the boundary.",[967],{"_967":2353},{"_272":640,"_139":1764},{"_939":152,"_940":152,"_941":152,"_942":447},"Read one gateway server revision, active or not.",{"_272":606,"_604":2372,"_602":2373,"_914":152},{"_884":76,"_885":71,"_886":73,"_887":2359,"_889":76,"_890":73,"_891":2180,"_882":462,"_893":79},"pl4h4qqenz",{"_894":2360,"_896":2361,"_898":2362,"_900":2363,"_902":2364,"_904":2365},{"_488":2371,"_490":73},{"_488":2370,"_490":79},{"_488":2369,"_490":76},{"_488":2368,"_490":71},{"_488":2367,"_490":76},{"_488":2366,"_490":73},"For a read-only single-entity fetch with full schema coverage and annotations carrying the safety profile, the description is adequate; the lack of an output schema means a note on the return shape would help, but the tool's shape is largely self-evident from its name and parameters.","Annotations already declare readOnlyHint=true, destructiveHint=false, openWorldHint=false, so the safety profile is covered. The description adds one genuinely useful behavioral fact — that inactive revisions are retrievable, not just active ones — but says nothing about error behavior when the revision does not exist.","A single short sentence with zero padding, and the operative qualifier ('active or not') is placed where it is read immediately. Nothing needs trimming and nothing is buried.","Schema description coverage is 100% and both parameters are documented inline, including the provenance of each value ('from list_gateway_server_revisions', 'from list_gateway_servers'). The description adds no parameter detail beyond the schema, so the baseline 3 applies.","'active or not' hints that the tool retrieves non-active revisions too, which is a mild usage cue, but there is no explicit when-to-use guidance and no alternatives named. An agent gets no direction on choosing this over get_gateway_server or list_gateway_server_revisions.","The description gives a specific verb ('Read') and resource ('gateway server revision') with a singular scope ('one'), which cleanly separates it from the sibling list_gateway_server_revisions. However, it never names or contrasts itself against any sibling tool (e.g. get_gateway_server), so it stops short of full differentiation.",[1038,1182],{"_1182":2374,"_1038":2375},{"_272":1187,"_139":2376},{"_272":640,"_139":1245},"The revision number, from list_gateway_server_revisions.",{"_939":152,"_940":152,"_941":152,"_942":447},"Read a gateway server's active revision: the configuration the gateway serves. Not found when no revision is active; use list_gateway_server_revisions then.",{"_272":606,"_604":2395,"_602":2396,"_914":152},{"_884":73,"_885":71,"_886":71,"_887":2382,"_889":76,"_890":71,"_891":988,"_882":467,"_893":71},"ht74tcx4p3",{"_894":2383,"_896":2384,"_898":2385,"_900":2386,"_902":2387,"_904":2388},{"_488":2394,"_490":71},{"_488":2393,"_490":71},{"_488":2392,"_490":76},{"_488":2391,"_490":71},{"_488":2390,"_490":73},{"_488":2389,"_490":71},"For a single-parameter read tool with no output schema, the description covers what is returned, the empty-state behavior, and the recovery path. Nothing needed to invoke it correctly is missing.","Annotations already declare readOnlyHint=true, destructiveHint=false, openWorldHint=false. The description adds real value beyond them by disclosing the 'not found' condition when no revision is active. It does not address the unusual idempotentHint=false or the shape of the returned configuration, leaving minor gaps.","Two short sentences, zero filler, and the core purpose is front-loaded ahead of the failure/alternative guidance.","Only one parameter, and the schema description coverage is 100% — the schema already explains server_slug and its origin in list_gateway_servers. The description adds nothing about the parameter, so the baseline 3 applies.","Explicitly gives the failure case ('Not found when no revision is active') and names the alternative to use in that case ('use list_gateway_server_revisions then'). This is exactly the when/when-not/alternative pattern.","States a specific verb (Read) and a precise resource (a gateway server's active revision) with the semantic gloss 'the configuration the gateway serves'. This clearly separates it from the sibling get_gateway_server_revision (a specific revision) and list_gateway_server_revisions.",[1038],{"_1038":2397},{"_272":640,"_139":1904},{"_939":152,"_940":152,"_941":152,"_942":447},"Read one escalation path revision, active or not, with its nodes and warnings.",{"_272":606,"_604":2416,"_602":2417,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":2403,"_889":76,"_890":73,"_891":688,"_882":467,"_893":76},"gm4kh3ox7y",{"_894":2404,"_896":2405,"_898":2406,"_900":2407,"_902":2408,"_904":2409},{"_488":2415,"_490":73},{"_488":2414,"_490":76},{"_488":2413,"_490":76},{"_488":2412,"_490":71},{"_488":2411,"_490":73},{"_488":2410,"_490":73},"With no output schema, the description compensates by naming the returned components (nodes, warnings). Safety is covered by annotations. It stops short of explaining failure modes (e.g., unknown revision) or what 'warnings' contains, but it is sufficient to call the tool correctly.","Annotations already cover the safety profile (readOnly=true, destructive=false), so the bar is lower. The description still adds value by disclosing that inactive revisions are readable and that the payload includes nodes and warnings — meaningful context for a tool with no output schema.","A single tight sentence with the verb and resource front-loaded, the active/inactive scope stated early, and the return contents at the end. No filler.","Schema description coverage is 100%, so both path_key and revision are already documented, including the pointer to list_escalation_path_revisions for revision provenance. The description adds no syntax or format detail beyond the schema, so the baseline of 3 applies.","Usage is implied rather than stated: 'one ... revision' and 'active or not' signal this is the fetch-a-specific-revision companion to list_escalation_path_revisions, but there is no explicit when-to-use or when-not-to-use guidance in the description itself.","States a specific verb (Read) and resource (one escalation path revision), and the qualifier 'active or not' distinguishes it from tools that only surface the current revision. It does not explicitly name siblings like get_escalation_path or list_escalation_path_revisions, but the scope is unambiguous.",[1094,1182],{"_1094":2418,"_1182":2419},{"_272":640,"_139":1971},{"_272":1187,"_139":2420},"The revision number, from list_escalation_path_revisions.",{"_939":152,"_940":152,"_941":152,"_942":447},"Read an escalation path's active revision: its nodes, working hours, repeat setting and warnings. Not found when the path has no active revision; use list_escalation_path_revisions then.",{"_272":606,"_604":2439,"_602":2440,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":2426,"_889":76,"_890":71,"_891":1075,"_882":467,"_893":71},"xa61lfmlh5",{"_894":2427,"_896":2428,"_898":2429,"_900":2430,"_902":2431,"_904":2432},{"_488":2438,"_490":71},{"_488":2437,"_490":71},{"_488":2436,"_490":76},{"_488":2435,"_490":71},{"_488":2434,"_490":73},{"_488":2433,"_490":73},"With no output schema, enumerating the returned fields is valuable, and the failure mode plus fallback makes the tool callable without surprises. Minor gaps remain (no statement of required access level or whether inactive/draft revisions are visible), but nothing essential is missing for a single-key read.","Annotations already declare readOnlyHint=true and destructiveHint=false, so safety is covered; the description adds a real behavioral fact beyond them - the tool errors when no active revision exists - plus what the response contains. It stops short of details like permission requirements or how 'active' resolution behaves when only drafts exist.","Two sentences, no filler: the first defines the read and its output, the second front-loads the error condition and the recovery path. Every clause carries information.","Only one parameter and schema description coverage is 100%, so the schema fully documents path_key. The description implies its meaning ('the escalation path's ... revision') but adds no syntax, format, or lookup guidance beyond what the schema provides; baseline 3 applies.","It states the exact failure condition ('Not found when the path has no active revision') and the alternative to use in that case ('use list_escalation_path_revisions then'). This is explicit when-to-use/when-not and names the fallback sibling, leaving nothing to inference.","The description names a specific verb ('Read') and resource ('escalation path's active revision'), then enumerates the payload (nodes, working hours, repeat setting, warnings). This distinguishes it from the sibling get_escalation_path_revision, which serves a revision-specific view, without the agent needing to open either schema.",[1094],{"_1094":2441},{"_272":640,"_139":1971},{"_939":152,"_940":152,"_941":152,"_942":447},"Read one audit event by id with its full data.",{"_272":606,"_604":2460,"_602":2461,"_914":152},{"_884":76,"_885":71,"_886":73,"_887":2447,"_889":76,"_890":73,"_891":2180,"_882":462,"_893":79},"kywlejz374",{"_894":2448,"_896":2449,"_898":2450,"_900":2451,"_902":2452,"_904":2453},{"_488":2459,"_490":73},{"_488":2458,"_490":79},{"_488":2457,"_490":76},{"_488":2456,"_490":71},{"_488":2455,"_490":76},{"_488":2454,"_490":73},"For a low-complexity single-parameter getter with a rich annotation set and no output schema, the description is nearly sufficient. It could only be improved by noting behavior on an unknown id or how the event is meant to be located.","Annotations already declare readOnlyHint=true, destructiveHint=false, openWorldHint=false, so the safety profile is fully covered. The description's only added behavioral signal is 'full data', which hints at completeness of the returned event but says nothing about missing-id errors or the odd idempotentHint=false.","One short sentence with the verb front-loaded and zero wasted words. The scoping detail ('with its full data') follows the identifying clause naturally.","Schema coverage is 100% and the single event_id parameter is documented there, so the schema does the heavy lifting. The description's 'by id' merely restates the parameter without adding format, source, or lookup semantics.","There is no explicit when-to-use guidance or mention of alternatives; the 'by id' phrasing only implies it is for a known event id. It never tells the agent to prefer list_audit_events when the id is unknown, nor states any prerequisites.","States a specific verb and resource ('Read one audit event by id') plus scope ('with its full data'), which cleanly separates it from the list_audit_events sibling. It does not explicitly name the sibling, so it falls just short of a 5.",[2462],{"_2462":2463},"event_id",{"_272":640,"_139":2464},"The audit event's id.",{"_939":152,"_940":152,"_941":152,"_942":447},"Read one agent: its kind and status, a page of its instances (id, name, owner, status, liveness, credentials) and a page of its provisioner tokens. Instance and token ids are what update_agent_instance, revoke_agent_credential and revoke_provisioner_token take.",{"_272":606,"_604":2483,"_602":2484,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":2470,"_889":76,"_890":71,"_891":1140,"_882":467,"_893":73},"f1i78uotw9",{"_894":2471,"_896":2472,"_898":2473,"_900":2474,"_902":2475,"_904":2476},{"_488":2482,"_490":71},{"_488":2481,"_490":73},{"_488":2480,"_490":76},{"_488":2479,"_490":71},{"_488":2478,"_490":73},{"_488":2477,"_490":73},"With no output schema, the description carries the return-shape burden and does a reasonable job naming the returned entities and their fields. It does not specify total counts, default ordering of instances, or error behavior for a missing slug, which is a minor residual gap.","Annotations already declare readOnlyHint=true, destructiveHint=false and openWorldHint=false, so safety is covered. The description adds real behavioral context beyond that: the pagination granularity of both collections (25 per page, tokens newest first) and the shape of what is returned, which matters because there is no output schema.","Two sentences, both front-loaded with the most decision-relevant information: what is read first, then the downstream id usage. No filler.","Schema description coverage is 100%, so every filter (instance_q, instance_status, the two offsets, provisioner_status) is already documented in the schema. The description adds no syntax or default detail for those filters, so the baseline of 3 applies.","Clearly positions the tool as the single-agent reader and explains that the returned instance and token ids are the inputs that update_agent_instance, revoke_agent_credential and revoke_provisioner_token take, which is genuine routing guidance. It stops short of stating when not to use it or naming list_agents as the multi-agent alternative explicitly.","States a specific verb and resource ('Read one agent') and then enumerates exactly what comes back: kind, status, a page of instances with named fields, and a page of provisioner tokens. An agent can distinguish this from list_agents purely from the description.",[1157],{"_1157":2485,"_2486":2487,"_2488":2489,"_2490":2491,"_2492":2493,"_2494":2495},{"_272":640,"_139":1159},"instance_q",{"_272":640,"_139":2500},"instance_offset",{"_272":1187,"_139":2499},"instance_status",{"_272":640,"_139":2498},"provisioner_offset",{"_272":1187,"_139":2497},"provisioner_status",{"_272":640,"_139":2496},"Only provisioner tokens that are active, expired or revoked.","How many matching provisioner tokens to skip; tokens come 25 to a page, newest first.","Only instances that are active or disabled.","How many matching instances to skip; instances come 25 to a page.","Only instances whose name contains this text.",{"_939":447,"_940":447,"_941":152,"_942":152},"Remove every credential the organization stores for one organization-held scope (its headers, or its OAuth token and client registration). Calls with the scope fail until it is connected again; the server and its revisions stay.",{"_272":606,"_604":2519,"_602":2520,"_914":152},{"_884":73,"_885":71,"_886":71,"_887":2506,"_889":73,"_890":71,"_891":988,"_882":467,"_893":73},"umuq7dj8fa",{"_894":2507,"_896":2508,"_898":2509,"_900":2510,"_902":2511,"_904":2512},{"_488":2518,"_490":71},{"_488":2517,"_490":73},{"_488":2516,"_490":73},{"_488":2515,"_490":71},{"_488":2514,"_490":73},{"_488":2513,"_490":71},"For a two-parameter destructive mutation with full annotation coverage and no output schema, the description supplies everything an agent needs: what is destroyed, what remains, and the runtime consequence. Nothing material is missing.","Annotations already declare destructive=true, idempotent=true, readOnlyHint=false, so the safety profile is covered. The description adds real value beyond that: it specifies the blast radius (all credentials for one scope), the consequence (calls fail until reconnected), and what is preserved (the server and its revisions). It does not mention permissions or return shape, but the key behavioral traits are disclosed.","Two tightly written sentences with zero filler. The removal target is front-loaded, and the survival/effect clause is appended for context without repetition.","Schema description coverage is 100%, so the baseline is 3 and both parameters (scope, server_slug) are already documented, including the list_gateway_servers pointer. The description goes slightly beyond by clarifying the nature of a scope (it holds headers or an OAuth token and client registration), which helps the caller understand what 'scope' identifies.","It conveys clear usage context: after this call, calls with the scope fail until it is reconnected, so an agent understands this is a disruptive, purposeful disconnect. It stops short of explicitly naming the reconnect/set alternatives (set_gateway_scope_headers, set_gateway_scope_oauth_client) or a when-not-to-use condition, so no exclusion is stated.","The description names a specific verb (remove credentials) and a precisely scoped resource (one organization-held scope), and enumerates exactly what is removed: headers, or OAuth token and client registration. It is clearly distinguishable from siblings like deactivate_gateway_server or set_gateway_scope_headers, which act on the server or set rather than removing credentials.",[1038,1237],{"_1237":2521,"_1038":2522},{"_272":640,"_139":2523},{"_272":640,"_139":1245},"The scope's name.",{"_939":152,"_940":447,"_941":152,"_942":152},"Stop serving a gateway server by deactivating its active revision. Its history and stored credentials are kept; activate_gateway_server_revision serves it again.",{"_272":606,"_604":2542,"_602":2543,"_914":152},{"_884":73,"_885":71,"_886":71,"_887":2529,"_889":76,"_890":71,"_891":892,"_882":467,"_893":73},"i4l6t8uqzf",{"_894":2530,"_896":2531,"_898":2532,"_900":2533,"_902":2534,"_904":2535},{"_488":2541,"_490":71},{"_488":2540,"_490":73},{"_488":2539,"_490":76},{"_488":2538,"_490":71},{"_488":2537,"_490":73},{"_488":2536,"_490":71},"With no output schema needed for a toggle, the description plus annotations cover everything an agent needs: what changes, what is preserved, and how to undo it. Nothing material is missing for a one-parameter tool.","Annotations already declare readOnly=false, destructive=false and idempotent=true; the description adds genuinely new behavioral facts — that history and stored credentials survive and that the operation is reversible via a named sibling. It does not mention auth requirements, but for a single-param toggle this is solid added context.","Two tight sentences: the action and scope first, then the state-preservation and reversal path. No filler, every clause carries information.","There is one parameter and schema coverage is 100%, with the schema itself explaining that the slug comes from list_gateway_servers and how it prefixes tool names. The description adds nothing about server_slug, so the baseline 3 applies.","Explicitly routes the agent to activate_gateway_server_revision for re-enabling, and the 'history and credentials are kept' line signals this is the soft-stop rather than the archiving path. It stops short of explicitly saying when this is preferable to archive_gateway_server, so it is clear context but not a full when/when-not.","States a specific verb+resource ('stop serving a gateway server by deactivating its active revision') and names the sibling that reverses it. The added note that history and credentials are kept implicitly separates it from archive_gateway_server and update_gateway_server.",[1038],{"_1038":2544},{"_272":640,"_139":1904},{"_939":152,"_940":152,"_941":152,"_942":152},"Create a team, optionally with its first members. Teams are escalation targets: a level that names a team reaches every current member, or runs the team's escalation policy when it has one. Requires a personal API key; organization keys cannot manage teams.",{"_272":606,"_604":2563,"_602":2564,"_914":152},{"_884":73,"_885":73,"_886":73,"_887":2550,"_889":76,"_890":71,"_891":1911,"_882":467,"_893":76},"ih24e1tsij",{"_894":2551,"_896":2552,"_898":2553,"_900":2554,"_902":2555,"_904":2556},{"_488":2562,"_490":71},{"_488":2561,"_490":76},{"_488":2560,"_490":76},{"_488":2559,"_490":73},{"_488":2558,"_490":73},{"_488":2557,"_490":73},"For a two-parameter creation tool with no output schema, the definition covers the action, the optional member placement, the domain role of teams, and the auth constraint. Nothing an agent needs to invoke it correctly is missing, though the return shape is unstated (acceptable absent an output schema).","Annotations already declare readOnlyHint=false, destructiveHint=false, and idempotentHint=false, so the safety profile is covered. The description adds genuinely non-derivable context: the personal-API-key requirement and the fact that organization keys cannot manage teams, which is a real operational constraint an agent must know before calling.","Three tight sentences, front-loaded with the action and its optional scope. The middle sentence on escalation semantics is arguably tangential to the act of creating a team, but it is short and earns its place as domain context.","Schema description coverage is 100%, so both parameters are already documented in the schema, and the baseline is 3. The phrase 'optionally with its first members' restates the optionality of membership_ids but adds no syntax, format, or sourcing detail beyond the schema.","It explains what a team IS ('escalation targets: a level that names a team reaches every current member...'), which implies the reason to create one, but it never states when to use create_team versus rename_team, set_team_members, or archive_team. Usage is implied rather than directed.","The description opens with a precise verb+resource ('Create a team') and immediately scopes it ('optionally with its first members'), which cleanly separates it from siblings like create_agent or create_escalation_path. An agent can identify the operation without opening the schema.",[68],{"_68":2565,"_1211":2566},{"_272":640,"_139":2570},{"_272":2567,"_926":2568,"_139":2569},[929,930],{"_272":640},"Membership ids (from list_members) to place on the team right away.","The team's name, unique among live teams.",{"_939":152,"_940":152,"_941":152,"_942":152},"Create a provisioner token for a service agent, which lets a fleet or platform create instances of the agent on its own, each with a credential of its own, through the Agent Approval Protocol. It cannot make approval requests itself. The answer carries the token exactly once: hand it to whoever runs the platform, because it cannot be read back.",{"_272":606,"_604":2589,"_602":2590,"_914":152},{"_884":71,"_885":71,"_886":71,"_887":2576,"_889":76,"_890":71,"_891":988,"_882":467,"_893":73},"r0bfbgvhs6",{"_894":2577,"_896":2578,"_898":2579,"_900":2580,"_902":2581,"_904":2582},{"_488":2588,"_490":71},{"_488":2587,"_490":73},{"_488":2586,"_490":76},{"_488":2585,"_490":71},{"_488":2584,"_490":71},{"_488":2583,"_490":71},"For a one-parameter creation tool with no output schema, the description covers what is created, its capabilities and limitations, and the one-time token return. Nothing essential for correct invocation is missing.","Annotations declare a non-read-only, non-idempotent, non-destructive write. The description adds critical operational facts beyond annotations: the token is returned exactly once, cannot be read back, and cannot itself make approval requests. These are essential behavioral details for correct handling.","Three sentences, front-loaded with purpose, then scope limitation, then the critical one-time token handling constraint. No filler; each sentence earns its place.","Schema coverage is 100% and the single required parameter agent_slug is fully documented in the schema. The description adds no additional parameter meaning, so the baseline of 3 applies.","Describes the scenario in which the tool is useful (a fleet or platform autonomously creating instances of the agent) but does not name alternatives or explicit when-not conditions. The context is clear, but no sibling alternatives are mentioned.","States a specific verb (Create) and resource (provisioner token) and explains the unique capability: letting a fleet or platform create agent instances with their own credentials via the Agent Approval Protocol. This clearly distinguishes it from sibling tools like create_agent_instance and create_agent.",[1157],{"_1157":2591},{"_272":640,"_139":1159},{"_939":152,"_940":152,"_941":152,"_942":152},"Create a pipeline that a delegate block in another pipeline can hand matching requests to, with its first revision active at once. Use it to give one team its own part of the approval flow: the entry pipeline decides which requests reach it, the new pipeline decides what happens to them. A person then grants the team pipeline.read, pipeline.write and pipeline.activate at pipeline: in the app so they can edit it and nothing else.",{"_272":606,"_604":2610,"_602":2611,"_914":152},{"_884":71,"_885":73,"_886":73,"_887":2597,"_889":76,"_890":71,"_891":892,"_882":467,"_893":73},"ju62g08fh6",{"_894":2598,"_896":2599,"_898":2600,"_900":2601,"_902":2602,"_904":2603},{"_488":2609,"_490":71},{"_488":2608,"_490":73},{"_488":2607,"_490":76},{"_488":2606,"_490":73},{"_488":2605,"_490":71},{"_488":2604,"_490":73},"For a creation tool with no output schema, the definition covers the immediate side effect (first revision active), key immutability, default escalation behavior, and the follow-up permission grant. Return shape is not described, but for a create operation whose effect is stated up front this is close to complete.","Annotations cover the safety profile (non-readonly, non-destructive, non-idempotent, closed-world), and the description adds genuinely new behavioral facts: the first revision goes live immediately, the key is permanent (archived pipelines included) and cannot be 'entry', and a person must grant pipeline.read/write/activate at pipeline:. These are exactly the operational details annotations cannot express.","Three sentences that are front-loaded with the core purpose, followed by usage and then the post-creation permission step. Dense but each sentence carries distinct load; only the mixed audience/workflow framing is slightly more elaborate than strictly necessary.","Schema description coverage is 100%, so the schema already documents all four parameters, including the omit/empty-list behavior for blocks and the escalation-path inheritance. The description reinforces the permission model around pipeline: but adds little parameter syntax or format detail beyond the schema, so the baseline 3 is correct.","Provides clear when-to-use context with the team-owns-part-of-the-approval-flow scenario and the split of responsibility between entry pipeline and new pipeline. It stops short of explicitly naming alternatives or exclusions (e.g., use update_pipeline to modify an existing pipeline), so it is strong context rather than a full routing guide.","States a specific verb (Create) and resource (a pipeline), and pins down exactly what the created object is for: a target a delegate block in another pipeline hands matching requests to. This is enough to distinguish it from update_pipeline, archive_pipeline, and preview_pipeline without opening any schema.",[967,68],{"_68":2612,"_964":2613,"_967":2614,"_971":2615},{"_272":640,"_139":2621},{"_272":2618,"_926":2619,"_139":2620},{"_272":640,"_139":2617},{"_272":640,"_139":2616},"Escalation path key for requests this pipeline hands to a person without a block naming a path, including requests that run out of it. Omit to inherit the defaults of the pipelines that delegate into it.","The new pipeline's key: lowercase letters, digits, dots, underscores and hyphens, up to 63 characters. It is taken for good, archived pipelines included, and can never be 'entry'.",[929,930],{"_272":606,"_914":447},"The first revision's blocks, in the shape update_pipeline takes. Omit or pass an empty list for a pipeline that passes every request that enters it to human review until it is edited.","The pipeline's display name, for example Payments.",{"_939":152,"_940":152,"_941":152,"_942":152},"Register an MCP server on the gateway with its first revision and activate it unless activate is false. Agents reach its tools as __, but only the tools set_agent_tool_access grants them, and every call still goes through the approval pipelines. Store an organization-held scope's credential afterwards with set_gateway_scope_headers; an OAuth scope is connected by a person in the app, and a member-held scope by each member on their Connections page.",{"_272":606,"_604":2640,"_602":2641,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":2627,"_889":76,"_890":71,"_891":1140,"_882":467,"_893":73},"m1gyq2biuv",{"_894":2628,"_896":2629,"_898":2630,"_900":2631,"_902":2632,"_904":2633},{"_488":2639,"_490":71},{"_488":2638,"_490":73},{"_488":2637,"_490":76},{"_488":2636,"_490":71},{"_488":2635,"_490":73},{"_488":2634,"_490":73},"For a create tool with no output schema and full annotation coverage, the description covers activation, downstream gating, and credential handling thoroughly. It does not indicate how to reference the newly created server (id/slug) for the follow-up calls it recommends, a minor gap.","Annotations declare non-read-only, non-destructive, non-idempotent, but the description adds substantial behavior: default activation, agent tool naming (__), access gating, the approval pipeline on every call, and the follow-up credential workflows per holder type. It stops short of error/failure behavior, but the added context is well beyond the annotations.","Three sentences, front-loaded with the core action, then agent-access implications, then credential follow-ups. Every sentence carries distinct, non-redundant information.","Schema description coverage is 100%, so the schema already documents url/tunnel_id exclusivity, scope structure, and slug rules in detail. The description restates activate's default and the slug naming, adding marginal value beyond the schema baseline.","Gives clear workflow context: activate defaults to true, tools are gated by set_agent_tool_access, and credentials are stored afterwards with set_gateway_scope_headers or via the app. It does not explicitly contrast with update_gateway_server or archive_gateway_server, but the create-vs-update distinction is unambiguous from the description.","States a specific verb and resource ('Register an MCP server on the gateway with its first revision') and distinguishes itself from sibling creation tools like create_pipeline and create_agent. The description makes clear this creates the server and its initial revision together.",[1038,68,1033],{"_92":2642,"_68":2643,"_1033":2644,"_921":2645,"_1036":2646,"_1038":2647},{"_272":640,"_139":1068},{"_272":640,"_139":1067},{"_272":2649,"_926":2650,"_139":1046},{"_272":2648,"_139":1043},{"_272":640,"_139":1041},{"_272":640,"_139":1040},[929,933],[929,930],{"_272":606,"_604":2651,"_602":2652,"_914":152},[68,139,1052,1050],{"_68":2653,"_1050":2654,"_1052":2655,"_139":2656,"_1055":2657,"_1057":2658},{"_272":640,"_139":1066},{"_272":640,"_139":1065},{"_272":640,"_139":1064},{"_272":640,"_139":1063},{"_272":2659,"_926":2660,"_139":1062},{"_272":640,"_139":1059},[929,930],{"_272":640},{"_939":152,"_940":152,"_941":152,"_942":152},"Create an escalation path with its first revision and activate it (unless activate is false). A path says who reviews a request the pipeline hands to a person: level nodes page targets (user targets by membership id, team targets by team id) and wait escalate_after before the next level; if_else nodes branch on urgency, working hours or the request itself; defer nodes hold the request until working hours. Call read_reference (escalation_paths) first. The key is permanent: pipelines name it, and it stays taken even after archiving.",{"_272":606,"_604":2679,"_602":2680,"_914":152},{"_884":71,"_885":73,"_886":73,"_887":2666,"_889":73,"_890":71,"_891":988,"_882":467,"_893":73},"oz0dhtqbz6",{"_894":2667,"_896":2668,"_898":2669,"_900":2670,"_902":2671,"_904":2672},{"_488":2678,"_490":71},{"_488":2677,"_490":73},{"_488":2676,"_490":73},{"_488":2675,"_490":73},{"_488":2674,"_490":71},{"_488":2673,"_490":73},"For a 6-parameter tool with nested objects and no output schema, the description covers node structure, activation, prerequisites, and key permanence. It does not mention required permissions or what the call returns, but those are minor omissions given the otherwise thorough treatment.","Annotations only cover the generic safety profile (not read-only, not destructive, not idempotent), so the description carries real weight here and delivers: the activation side effect and its default, the permanence of path_key ('stays taken even after archiving'), and the semantics of level/if_else/defer nodes and repeat behavior. This is context an agent could not infer from the annotations.","Front-loaded with the action and its activation behavior, followed by domain context and the read_reference prerequisite. Four dense sentences with little waste, though the middle node-explanation sentence is long and partially duplicates the nodes parameter description.","Schema coverage is 100%, so the baseline is 3, but the description adds meaning beyond the schema: node-type behavior (paging targets by membership id vs team id, escalate_after waiting), the activate default and its consequence, and the permanent, non-reusable nature of path_key. The extra framing is genuinely useful for constructing valid nodes.","Gives an explicit prerequisite: 'Call read_reference (escalation_paths) first.' It does not spell out when to choose create versus update_escalation_path, but the create semantics are unambiguous enough that the gap is small.","States a specific verb and resource ('Create an escalation path with its first revision') and immediately clarifies the domain concept ('A path says who reviews a request the pipeline hands to a person'). This cleanly separates it from update_escalation_path, archive_escalation_path, and the activate_*_revision siblings.",[1094,68,915],{"_68":2681,"_915":2682,"_917":2683,"_921":2684,"_1094":2685,"_923":2686},{"_272":640,"_139":1106},{"_272":2690,"_926":2691,"_139":1105},{"_272":606,"_139":935,"_914":447},{"_272":2689,"_139":1102},{"_272":640,"_139":1100},{"_272":2687,"_926":2688,"_139":1099},[929,930],{"_272":606,"_914":447},[929,933],[929,930],{"_272":606,"_914":447},{"_939":152,"_940":152,"_941":152,"_942":152},"Register an instance of a service agent and mint its credential, for a runtime the withhuman CLI does not install (a service of yours speaking the Agent Approval Protocol). The answer carries the credential exactly once: hand it to whoever configures that runtime, because it cannot be read back. Personal agents' instances belong to the member who enrolls them through the CLI.",{"_272":606,"_604":2710,"_602":2711,"_914":152},{"_884":73,"_885":73,"_886":73,"_887":2697,"_889":76,"_890":71,"_891":1281,"_882":467,"_893":73},"ibqbsrh9nf",{"_894":2698,"_896":2699,"_898":2700,"_900":2701,"_902":2702,"_904":2703},{"_488":2709,"_490":71},{"_488":2708,"_490":73},{"_488":2707,"_490":76},{"_488":2706,"_490":73},{"_488":2705,"_490":73},{"_488":2704,"_490":73},"With no output schema, the description correctly fills the gap by explaining the returned credential and its one-time visibility, plus the scope limitation. It leaves minor gaps around error cases and whether repeated calls create duplicate instances, but covers what an agent needs to invoke it correctly.","Annotations cover the safety profile (non-read-only, non-destructive, non-idempotent, closed-world), and the description adds the critical trait that the credential is returned exactly once and cannot be read back. That is actionable behavior beyond the annotations, though it says nothing about failure or duplicate-instance behavior.","Three sentences, each carrying distinct information (what it does, the one-time credential, the personal-agent exclusion), with the core action front-loaded. Slightly dense but no wasted sentences.","Schema description coverage is 100%, so name, metadata, and agent_slug are already documented in the schema. The description confirms the agent_slug source but adds no syntax or format guidance beyond what the schema provides; baseline 3 applies.","Gives a clear usage condition (\"for a runtime the withhuman CLI does not install\") and an explicit exclusion: personal agents' instances belong to the member who enrolls them through the CLI. It routes the agent away from misuse, though it does not name a sibling tool as the alternative.","States a specific verb and resource (\"Register an instance of a service agent and mint its credential\") and scopes it to a runtime the CLI cannot install, distinguishing it from create_agent and update_agent_instance. An agent can identify the operation without opening the schema.",[1157,68],{"_68":2712,"_2713":2714,"_1157":2715},{"_272":640,"_139":2717},"metadata",{"_272":606,"_139":2716,"_914":447},{"_272":640,"_139":1159},"Any JSON object to store with the instance, such as a region or pod name.","A name for the instance, unique within the agent. Reviewers see it next to the agent's name.",{"_939":152,"_940":152,"_941":152,"_942":152},"Create an agent. Nothing reaches it until a runtime is connected: with the withhuman CLI on the machine the agent runs on, or with create_agent_instance or create_provisioner_token for a service you run. Every request it makes starts in the entry pipeline; give it rules with a branch on /agent/slug there, or a delegate to a pipeline of its own.",{"_272":606,"_604":2736,"_602":2737,"_914":152},{"_884":73,"_885":71,"_886":71,"_887":2723,"_889":76,"_890":73,"_891":1281,"_882":467,"_893":73},"ydyq25z0it",{"_894":2724,"_896":2725,"_898":2726,"_900":2727,"_902":2728,"_904":2729},{"_488":2735,"_490":73},{"_488":2734,"_490":73},{"_488":2733,"_490":76},{"_488":2732,"_490":71},{"_488":2731,"_490":73},{"_488":2730,"_490":71},"For a mutation tool with no output schema, the description covers the essential post-creation mental model (runtime connection and entry-pipeline rules). Combined with the annotations carrying the safety profile and the schema carrying full parameter docs, nothing critical is missing.","Annotations already declare the mutation profile (readOnlyHint=false, destructiveHint=false, idempotentHint=false), so the safety bar is covered. The description adds genuinely non-obvious behavioral context beyond them: the agent receives nothing until a runtime is connected, and every request starts in the entry pipeline with branching on /agent/slug.","Front-loaded with the imperative verb+resource, then dense but non-redundant sentences on runtime and pipeline setup. No filler; every sentence carries actionable information.","Schema description coverage is 100%, so kind, name and slug are already fully documented in the schema, including the service/personal distinction and the slug format rules. The description adds no parameter-level detail beyond the schema, so the baseline of 3 applies.","Provides clear workflow context: it explains that a created agent is inert until a runtime is connected and names the two paths (withhuman CLI, or create_agent_instance/create_provisioner_token). This tells an agent where this tool sits in the sequence, though it never states an explicit when-not-to-use condition.","States a specific verb and resource (\"Create an agent\") that is immediately distinguishable from read/update/archive siblings. It names related tools (create_agent_instance, create_provisioner_token) but doesn't explicitly contrast create_agent with update_agent or the instance tools, leaving full sibling differentiation to inference.",[24,68],{"_2738":2739,"_68":2740,"_24":2741},"kind",{"_272":640,"_139":2744},{"_272":640,"_139":2743},{"_272":640,"_139":2742},"Permanent identifier: lowercase letters, digits, dots, underscores or dashes, unique for good (for example support-assistant).","Display name, unique among live agents.","service (the default): an agent the organization runs, whose instances belong to nobody. personal: one people run for themselves, such as a coding assistant; each instance belongs to the member who enrolled it. Fixed for good.",{"_939":152,"_940":447,"_941":152,"_942":152},"Archive a webhook endpoint by deactivating its active revision. Its history and signing secret are kept, and activate_webhook_endpoint_revision brings it back. Refused while an active pipeline revision posts to it (list_webhook_endpoint_uses names them).",{"_272":606,"_604":2763,"_602":2764,"_914":152},{"_884":73,"_885":71,"_886":71,"_887":2750,"_889":76,"_890":71,"_891":988,"_882":467,"_893":71},"iquhumwthd",{"_894":2751,"_896":2752,"_898":2753,"_900":2754,"_902":2755,"_904":2756},{"_488":2762,"_490":71},{"_488":2761,"_490":71},{"_488":2760,"_490":76},{"_488":2759,"_490":71},{"_488":2758,"_490":73},{"_488":2757,"_490":71},"With no output schema, the description still covers the essentials for a single-param mutation: what changes, what is preserved, how to undo it, and when the call will fail. Nothing needed to invoke it correctly is absent.","Annotations cover the safety profile (idempotent=true, destructive=false, not read-only), but the description adds genuinely new behavior: history and signing secret are retained, the action is reversible, and there is a server-side refusal precondition with a diagnostic tool. What is missing is auth/permission requirements, but the added context is substantial beyond the structured fields.","Three compact sentences, no waste, and the core action is front-loaded ahead of the retention/reversal and refusal details. Every sentence carries distinct information an agent needs.","Schema description coverage is 100% for the single endpoint_key parameter, so the schema already documents it. The description adds no format, syntax, or lookup detail beyond 'the webhook endpoint's key', making the baseline 3 appropriate.","It explicitly names the reverse operation ('activate_webhook_endpoint_revision brings it back') and the condition under which the call is refused ('while an active pipeline revision posts to it'), pointing to list_webhook_endpoint_uses for diagnosis. Both when-to-use and when-not-to-use are given, with a named alternative.","The description states a specific verb and resource ('Archive a webhook endpoint') and clarifies the mechanism ('by deactivating its active revision'), which cleanly separates it from the activate_* sibling. An agent can distinguish it from the other 70+ archive_*/activate_* tools without opening the schema.",[1184],{"_1184":2765},{"_272":640,"_139":1186},{"_939":152,"_940":447,"_941":152,"_942":152},"Deactivate a team's escalation policy, so a path level targeting the team reaches every member at once again. Its revisions stay and can be activated later.",{"_272":606,"_604":2784,"_602":2785,"_914":152},{"_884":73,"_885":71,"_886":71,"_887":2771,"_889":76,"_890":73,"_891":1911,"_882":467,"_893":76},"e702bzj6gg",{"_894":2772,"_896":2773,"_898":2774,"_900":2775,"_902":2776,"_904":2777},{"_488":2783,"_490":73},{"_488":2782,"_490":76},{"_488":2781,"_490":76},{"_488":2780,"_490":71},{"_488":2779,"_490":73},{"_488":2778,"_490":71},"For a single-parameter tool with no output schema and annotations covering the safety profile, the description supplies the two things an agent most needs: what changes operationally and that the change is recoverable. Nothing required to call it correctly is missing.","Annotations already declare readOnlyHint=false, destructiveHint=false, and idempotentHint=true, so the safety profile is partly covered. The description adds genuinely useful context beyond that: revisions are retained and the policy can be activated later, confirming the operation is reversible rather than a hard delete. No permissions or error behavior is described, keeping it out of the 5 range.","Two short sentences with the core action front-loaded and no redundant restatement of the name or schema. Every clause carries distinct information (the routing effect, then the revision retention).","There is a single parameter with 100% schema description coverage, and the schema already tells the agent that team_id comes from list_teams. The description refers to 'a team's escalation policy' but adds no format, source, or constraint detail beyond the schema, so the baseline of 3 applies.","Usage is only implied: the reader infers you archive when you want to revert to default team-level routing rather than reconfigure the policy. There is no explicit when-to-use, when-not-to-use, or named alternative (e.g., why archive instead of update, or how this relates to activate_team_escalation_policy_revision).","The description gives a specific verb and resource ('Deactivate a team's escalation policy') and explains the operational effect — restoring team-level fan-out routing. It implicitly distinguishes itself from update_team_escalation_policy and activate_team_escalation_policy_revision by noting that revisions survive and can be reactivated, but it never names those alternatives outright.",[919],{"_919":2786},{"_272":640,"_139":934},{"_939":447,"_940":447,"_941":152,"_942":152},"Archive a team: it leaves routing, its own and team-scoped grants are revoked and its escalation policy is deactivated; its roster and history stay readable and its name can be reused. Refused while an active escalation path or a pending request targets it, and for a team your directory manages. Requires a personal API key.",{"_272":606,"_604":2805,"_602":2806,"_914":152},{"_884":71,"_885":71,"_886":71,"_887":2792,"_889":76,"_890":71,"_891":988,"_882":467,"_893":73},"cjdirl54zz",{"_894":2793,"_896":2794,"_898":2795,"_900":2796,"_902":2797,"_904":2798},{"_488":2804,"_490":71},{"_488":2803,"_490":73},{"_488":2802,"_490":76},{"_488":2801,"_490":71},{"_488":2800,"_490":71},{"_488":2799,"_490":71},"With no output schema, one fully-documented parameter, and annotations covering the safety profile, the description supplies everything an agent needs: effects, side effects, refusal conditions, and credential requirement.","Although annotations already flag this as destructive and idempotent, the description adds substantial context the annotations cannot convey: what exactly is revoked/deactivated, that the roster and history remain readable, that the name becomes reusable, the refusal preconditions, and the personal-API-key requirement.","Three sentences, each front-loaded with the highest-value information (effect, then preconditions, then auth). Every clause adds a distinct fact with no filler.","Schema coverage is 100% and the single parameter is documented as 'The team's id, from list_teams.' The description adds no further parameter syntax or format detail, so the baseline of 3 applies with the schema carrying the load.","It states explicit refusal conditions (an active escalation path or pending request targeting the team, or a team managed by your directory) which tell the agent when the call will fail, plus the auth requirement. It does not name a specific alternative tool to use instead, so it stops short of full routing guidance.","The description opens with a specific verb+resource ('Archive a team') and then enumerates the concrete consequences: leaving routing, revoking own and team-scoped grants, deactivating the escalation policy. This clearly separates it from siblings like rename_team, update_team, or archive_team_escalation_policy.",[919],{"_919":2807},{"_272":640,"_139":934},{"_939":447,"_940":447,"_941":152,"_942":152},"Archive a pipeline: its revisions stay readable but none is active, its key is taken for good, and no request can reach it. Refused for the entry pipeline and while any active revision still delegates to it (list_pipeline_uses names them; remove those delegate blocks first).",{"_272":606,"_604":2827,"_602":2828,"_914":152},{"_884":71,"_885":71,"_886":71,"_887":2813,"_889":76,"_890":71,"_891":2814,"_882":467,"_893":71},"rps3o42n8b",{"_894":2815,"_896":2816,"_898":2817,"_900":2818,"_902":2819,"_904":2820},4.7,{"_488":2826,"_490":71},{"_488":2825,"_490":71},{"_488":2824,"_490":76},{"_488":2823,"_490":71},{"_488":2822,"_490":71},{"_488":2821,"_490":71},"No output schema exists, yet the description covers outcome, irreversibility, post-conditions, and preconditions. Combined with annotations that carry the safety profile, an agent has everything needed to decide and call correctly.","Annotations already flag destructive=true and idempotent=true, but the description adds the concrete blast radius: revisions survive in readable form, nothing stays active, the key is irrecoverably consumed, and requests can no longer reach it. It also discloses the two refusal preconditions, which annotations cannot express.","Two dense sentences with the primary effect stated first, followed by the refusal conditions and the remedy. No filler clauses; each phrase carries distinct operational information.","There is a single parameter at 100% schema description coverage, and the schema already explains the 'entry' key versus a delegate target key. The description reinforces the entry-pipeline concept only indirectly through the refusal rule and adds no format or syntax detail, so the baseline 3 applies.","Explicitly states when the call is refused (the entry pipeline, or while an active revision still delegates to it) and points to the diagnostic sibling list_pipeline_uses to identify blockers. It also gives the remediation step: remove the delegate blocks first.","Names a specific verb (archive) and resource (pipeline), then spells out the three concrete effects: revisions remain readable, the key is permanently consumed, and no request can reach it. This clearly separates it from the many sibling archive_* tools, which target agents, teams, webhooks, etc.",[967],{"_967":2829},{"_272":640,"_139":1764},{"_939":447,"_940":447,"_941":152,"_942":152},"Archive a gateway server for good: it is never served again, takes no new revisions and its slug cannot be reused, and every stored credential (the organization's and every member's) is dropped. Its history stays readable. To pause a server instead, use deactivate_gateway_server.",{"_272":606,"_604":2848,"_602":2849,"_914":152},{"_884":71,"_885":71,"_886":71,"_887":2835,"_889":76,"_890":71,"_891":2814,"_882":467,"_893":71},"nn39idjrtz",{"_894":2836,"_896":2837,"_898":2838,"_900":2839,"_902":2840,"_904":2841},{"_488":2847,"_490":71},{"_488":2846,"_490":71},{"_488":2845,"_490":76},{"_488":2844,"_490":71},{"_488":2843,"_490":71},{"_488":2842,"_490":71},"For a one-parameter, irreversible mutation with no output schema, the description supplies the destructive effects, the non-reuse constraint, credential handling, and post-state readability, while annotations cover the safety profile. Nothing an agent needs to call it correctly is missing.","Annotations already declare destructiveHint=true and idempotentHint=true, but the description goes well beyond that with concrete consequences: the slug cannot be reused, every org and member credential is dropped, and history remains readable. That is exactly the extra destructive-effect detail an agent needs before acting.","Two sentences, no filler, with the irreversible consequences front-loaded and the alternative routed in the closing clause. Every clause earns its place.","Schema description coverage is 100% and the single parameter is fully documented in the schema, including where to get the slug and how it prefixes tool names. The description adds no additional parameter semantics, so the baseline 3 applies.","Explicitly names the alternative and the condition selecting it: 'To pause a server instead, use deactivate_gateway_server.' Nothing is left to inference about when to pick pause vs. permanent archive.","States a specific verb+resource ('Archive a gateway server') and immediately clarifies irreversible scope: never served again, no new revisions, slug not reusable, credentials dropped. This distinguishes it cleanly from the sibling deactivate_gateway_server and from archive_agent/archive_pipeline.",[1038],{"_1038":2850},{"_272":640,"_139":1904},{"_939":447,"_940":447,"_941":152,"_942":152},"Archive an escalation path for good: it stops being active, refuses new revisions, and keeps its key and history. Refused while any active pipeline revision names it (list_escalation_path_uses names them); update those pipelines first.",{"_272":606,"_604":2869,"_602":2870,"_914":152},{"_884":71,"_885":71,"_886":71,"_887":2856,"_889":76,"_890":71,"_891":988,"_882":467,"_893":73},"uiyjvjz47g",{"_894":2857,"_896":2858,"_898":2859,"_900":2860,"_902":2861,"_904":2862},{"_488":2868,"_490":71},{"_488":2867,"_490":73},{"_488":2866,"_490":76},{"_488":2865,"_490":71},{"_488":2864,"_490":71},{"_488":2863,"_490":71},"For a one-parameter destructive mutation with no output schema, the description covers what changes, what is preserved, what causes failure, and how to unblock. Nothing needed to call it correctly is missing.","With annotations only covering the safety profile (destructive/idempotent/not-read-only), the description supplies the missing substance: the path stops being active, refuses new revisions, and retains its key and history. It also discloses the refusal condition that makes the call fail, which an agent cannot infer from the annotations.","Two tight sentences with no filler; the irreversible nature and the retention of key/history are front-loaded before the failure precondition. Every clause earns its place.","Schema coverage is 100% and there is a single path_key parameter already documented as 'The escalation path's key.' The description adds no syntax, format, or lookup guidance beyond that, so the baseline of 3 applies.","Gives a clear blocking condition ('refused while any active pipeline revision names it') and names the diagnostic helper to run first (list_escalation_path_uses) plus the resolution step ('update those pipelines first'). It is precondition guidance rather than an explicit use-this-vs-that comparison, so it falls just short of a 5.","States a specific verb+resource (archive an escalation path) and immediately scopes the semantics with 'for good', distinguishing it from the deactivate/revoke siblings in the tool list. An agent can tell what this does without opening the schema.",[1094],{"_1094":2871},{"_272":640,"_139":1971},{"_939":447,"_940":447,"_941":152,"_942":152},"Archive an agent: every live credential and provisioner token is revoked for good, no instance can be enrolled under it, and its name is freed while its slug stays taken. Instances and history stay readable. restore_agent brings the agent back, but revoked credentials stay revoked.",{"_272":606,"_604":2890,"_602":2891,"_914":152},{"_884":71,"_885":71,"_886":71,"_887":2877,"_889":73,"_890":71,"_891":2814,"_882":467,"_893":73},"qxw2ax35rj",{"_894":2878,"_896":2879,"_898":2880,"_900":2881,"_902":2882,"_904":2883},{"_488":2889,"_490":71},{"_488":2888,"_490":73},{"_488":2887,"_490":73},{"_488":2886,"_490":71},{"_488":2885,"_490":71},{"_488":2884,"_490":71},"For a one-parameter destructive mutation with no output schema, the description covers effects, reversibility, and the fate of dependent resources and credentials. Nothing material to correct invocation is missing, and annotations already cover the safety profile.","Goes well beyond the annotations (destructiveHint=true, idempotentHint=true) by detailing exactly what is destroyed (live credentials and provisioner tokens, permanently), what is preserved (instances and history remain readable), and the asymmetric name/slug outcome. It also warns that restoration is partial, which is the single most important behavioral fact an agent caller needs.","Three tight sentences with no filler, front-loaded with the main action and its irreversible consequences, then the preservation guarantees, then the restore caveat. Each sentence carries information an agent cannot get from the schema or annotations.","Schema coverage is 100% and there is a single parameter, so the baseline is 3; the description nevertheless adds genuine meaning by explaining that the slug stays reserved after archival, which affects how agent_slug resolves and how callers should plan re-creation.","The description makes the use context clear (permanent teardown of an agent) and explicitly names restore_agent as the inverse operation, along with the crucial caveat that restore does not un-revoke credentials. It does not spell out when-not to archive (e.g., prefer deactivation/update_agent for temporary pauses), so it stops short of full alternative routing.","States a specific verb and resource (archive_agent) and immediately enumerates the scope of the operation: credential/provisioner-token revocation, enrollment lockout, name freed but slug retained. This distinguishes it clearly from siblings like archive_team, archive_pipeline, and archive_gateway_server, which do not carry these agent-specific effects.",[1157],{"_1157":2892},{"_272":640,"_139":1159},{"_939":152,"_940":447,"_941":152,"_942":152},"Make an existing webhook endpoint revision the active one (a rollback, activating a draft, or bringing an archived endpoint back). Webhook blocks post to the active revision's URL from their next delivery on.",{"_272":606,"_604":2911,"_602":2912,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":2898,"_889":76,"_890":71,"_891":1140,"_882":467,"_893":73},"wsiygrk0wc",{"_894":2899,"_896":2900,"_898":2901,"_900":2902,"_902":2903,"_904":2904},{"_488":2910,"_490":71},{"_488":2909,"_490":73},{"_488":2908,"_490":76},{"_488":2907,"_490":71},{"_488":2906,"_490":73},{"_488":2905,"_490":73},"For a two-parameter mutation with full annotation coverage and no output schema, the description supplies what the agent needs to act correctly and predict consequences. The only minor gap is not stating explicitly what happens to the previously active revision (demoted to draft/archived?), though 'rollback' implies it.","Annotations already declare the mutation profile (readOnlyHint=false, idempotentHint=true, destructiveHint=false), so the bar is lower. The description still adds genuinely non-obvious behavior: webhook blocks begin posting to the newly active revision's URL from their next delivery, and archived endpoints can be reactivated — neither is derivable from the annotations or schema.","Two sentences, no redundancy. The core action is front-loaded and the parenthetical clarifies scope economically; the second sentence carries the delivery-timing behavior without padding.","Schema description coverage is 100%; both parameters (endpoint_key, revision) are already documented, including the pointer to list_webhook_endpoint_revisions for the revision number. The description adds no syntax or format detail beyond the schema, so the baseline 3 applies.","Gives clear context for when the tool applies (rolling back, promoting a draft, restoring an archived endpoint), which tells the agent what situations call for activation. It stops short of naming alternatives or prerequisites, but the resource-specific name makes the choice among siblings unambiguous.","States a specific verb+resource ('make an existing webhook endpoint revision the active one') and immediately disambiguates the word 'activate' by enumerating the three scenarios it covers: rollback, draft activation, and unarchiving. The resource name differentiates it from the activate_*_revision siblings (pipeline, gateway_server, escalation_path, team_escalation_policy) without needing to name them.",[1184,1182],{"_1182":2913,"_1184":2914},{"_272":1187,"_139":2152},{"_272":640,"_139":1186},{"_939":152,"_940":447,"_941":152,"_942":152},"Make an existing team escalation policy revision the active one (a rollback, or activating a draft). Refused when it names someone who has since left the team. Requests already routed keep the revision they pinned.",{"_272":606,"_604":2933,"_602":2934,"_914":152},{"_884":73,"_885":71,"_886":71,"_887":2920,"_889":76,"_890":71,"_891":892,"_882":467,"_893":73},"qkiqt1uka1",{"_894":2921,"_896":2922,"_898":2923,"_900":2924,"_902":2925,"_904":2926},{"_488":2932,"_490":71},{"_488":2931,"_490":73},{"_488":2930,"_490":76},{"_488":2929,"_490":71},{"_488":2928,"_490":73},{"_488":2927,"_490":71},"With no output schema, the description carries the relevant outcome information (routed requests keep their pinned revision) and the failure case. An agent has everything needed to invoke this correctly.","Annotations already declare non-read-only, idempotent, non-destructive, so the safety profile is covered. The description adds genuinely new behavioral context beyond them: a precondition-failure mode and the scoping of effects on already-routed requests. It stops short of describing auth requirements or the response shape.","Three tight sentences: the action first, then the failure condition, then the side-effect scope. Every sentence carries distinct information with no filler.","Schema coverage is 100%, with both parameters documented (team_id points to list_teams, revision points to list_team_escalation_policy_revisions). The description adds no syntax or format detail beyond the schema, so the baseline 3 is appropriate.","Explains the two when-to-use cases (rollback, draft activation) and gives a concrete exclusion rule ('refused when it names someone who has since left the team'). It does not explicitly name a sibling alternative or prerequisite lookup tool, keeping it just below a 5.","States a specific verb (make active) plus the exact resource (team escalation policy revision), and clarifies the two triggering scenarios (rollback or activating a draft). This cleanly separates it from the sibling activate_* tools for escalation paths, gateway servers, pipelines, and webhook endpoints.",[919,1182],{"_919":2935,"_1182":2936},{"_272":640,"_139":934},{"_272":1187,"_139":2221},{"_939":152,"_940":447,"_941":152,"_942":152},"Make an existing pipeline revision the active one (a rollback, or activating a draft). New requests use it immediately; requests already in flight keep the revision they started on.",{"_272":606,"_604":2955,"_602":2956,"_914":152},{"_884":73,"_885":71,"_886":73,"_887":2942,"_889":76,"_890":73,"_891":73,"_882":467,"_893":73},"eho6qdp37b",{"_894":2943,"_896":2944,"_898":2945,"_900":2946,"_902":2947,"_904":2948},{"_488":2954,"_490":73},{"_488":2953,"_490":73},{"_488":2952,"_490":76},{"_488":2951,"_490":71},{"_488":2950,"_490":73},{"_488":2949,"_490":73},"For a two-parameter mutation with no output schema, the definition covers purpose, triggers, and request-handling semantics, with annotations carrying the safety profile. Only prerequisites and error behavior are absent, which is a minor gap rather than a blocking one.","Annotations already establish write/idempotent/non-destructive, so the bar is lower, and the description adds substantive transition semantics: new requests pick up the revision immediately while in-flight requests stay pinned to their original revision. That is exactly the kind of post-condition an agent cannot infer from annotations. It does not mention permissions or failure modes.","Two sentences, no filler. The core action is front-loaded and the transition behavior follows as qualification.","Schema description coverage is 100%, and the schema itself documents both parameters including a pointer to list_pipeline_revisions for valid revision numbers. The description adds no parameter-level meaning beyond the schema, so the baseline 3 applies.","It states the two conditions under which the tool is the right choice ('a rollback, or activating a draft'), which is real routing guidance. It gives no when-not-to-use guidance or prerequisites (e.g. whether the revision must be in draft, what happens if it is already active).","The description gives a precise verb+resource ('Make an existing pipeline revision the active one') and immediately qualifies the two scenarios that trigger it — rollback or activating a draft. It is unambiguous against list_pipeline_revisions and update_pipeline, though it never explicitly contrasts itself with sibling activation tools like activate_webhook_endpoint_revision.",[967,1182],{"_1182":2957,"_967":2958},{"_272":1187,"_139":2310},{"_272":640,"_139":2309},{"_939":152,"_940":447,"_941":152,"_942":152},"Make an existing gateway server revision the active one (a rollback, activating a draft, or serving a deactivated server again). The gateway serves it from its next configuration poll; scopes the revision leaves out are retired and their credentials dropped.",{"_272":606,"_604":2977,"_602":2978,"_914":152},{"_884":71,"_885":71,"_886":73,"_887":2964,"_889":76,"_890":71,"_891":1075,"_882":467,"_893":73},"xf9vc7ugji",{"_894":2965,"_896":2966,"_898":2967,"_900":2968,"_902":2969,"_904":2970},{"_488":2976,"_490":71},{"_488":2975,"_490":73},{"_488":2974,"_490":76},{"_488":2973,"_490":71},{"_488":2972,"_490":71},{"_488":2971,"_490":73},"For an idempotent mutation with no output schema, the description covers action, timing, and side effects, which is nearly everything needed to call it correctly. It does not touch failure modes (unknown revision, invalid slug) or confirmation requirements, so it is slightly short of complete.","Annotations only declare the safety profile (readOnly=false, idempotent=true, destructive=false); the description supplies the non-obvious operational behavior: activation takes effect at the gateway's next configuration poll, and scopes omitted by the revision are retired and their credentials dropped. Disclosing that credential-dropping side effect is exactly the kind of context an agent cannot infer from structured fields.","Two tight sentences with zero filler; the action and the effect/timing are front-loaded before the side-effect caveat. Every clause carries information the agent needs.","Schema description coverage is 100% and both parameters already point at their source tools (list_gateway_server_revisions, list_gateway_servers). The description adds no syntax, format, or constraint detail beyond that, so the baseline 3 applies.","Gives clear context for when to reach for it (rollback, draft promotion, reviving a deactivated server), which is explicit usage guidance. It does not name an alternative or state when-not to use it (e.g., versus deactivate_gateway_server or update_gateway_server), so it stops short of full routing.","States a specific verb and resource ('Make an existing gateway server revision the active one') and immediately parenthesizes the concrete scenarios it covers: rollback, activating a draft, re-serving a deactivated server. The activate_* siblings all operate on different resources, and the name plus description make the resource unambiguous without opening any schema.",[1038,1182],{"_1182":2979,"_1038":2980},{"_272":1187,"_139":2376},{"_272":640,"_139":1245},{"_939":152,"_940":447,"_941":152,"_942":152},"Make an existing escalation path revision the active one (a rollback, or activating a draft).",{"_272":606,"_604":2999,"_602":3000,"_914":152},{"_884":76,"_885":71,"_886":73,"_887":2986,"_889":76,"_890":73,"_891":474,"_882":467,"_893":76},"mup9ou6xay",{"_894":2987,"_896":2988,"_898":2989,"_900":2990,"_902":2991,"_904":2992},{"_488":2998,"_490":73},{"_488":2997,"_490":76},{"_488":2996,"_490":76},{"_488":2995,"_490":71},{"_488":2994,"_490":76},{"_488":2993,"_490":73},"With only two well-documented required parameters, no nested objects, no output schema, and annotations covering idempotency and safety, very little is left for the description to carry. The one remaining gap is the side effect on the previously active revision.","Annotations already declare readOnlyHint=false, idempotentHint=true, destructiveHint=false and openWorldHint=false, so the safety profile is covered. The description usefully adds that a prior revision is being superseded ('rollback'), but never says what happens to the previously active revision or whether the change can be undone — the main behavioral question for an activation tool.","A single front-loaded sentence with zero filler; the core action leads and the qualifier follows. Nothing to trim.","Schema description coverage is 100%, and the schema already explains both path_key and revision (including that the revision number comes from list_escalation_path_revisions). The description adds nothing about parameter format or validation, so the baseline of 3 applies.","'a rollback, or activating a draft' implies when the tool is appropriate, but there is no explicit when-not guidance, no mention of prerequisites (e.g. revision must exist / come from list_escalation_path_revisions), and no indication of what to use instead for other lifecycle actions.","States a specific verb ('Make ... the active one') and resource ('escalation path revision'), and the parenthetical clarifies the two scenarios it covers: rollback and draft activation. It cleanly distinguishes this resource's revision-activation from the many activate_*_revision siblings, though it doesn't explicitly name an alternative tool.",[1094,1182],{"_1094":3001,"_1182":3002},{"_272":640,"_139":1971},{"_272":1187,"_139":2420},"challengeToken","domains",[3007],"githubNamespaceLogin",{"_3008":3009,"_3010":3011,"_3012":3013,"_2738":3014},"dnsRecordName","_glama-claim.app.withhuman.ai","domain","app.withhuman.ai","httpUrl","https://app.withhuman.ai/.well-known/glama.json","host","pageSessionReferenceId","01a13f50-1ea0-7520-b9b5-1eafc4aa0384","renderedAt",1792063577759,"toasts",[],"visitor",{"_3023":3024,"_3025":3026},"viewportWidth",1280,"visitorSession",{"_544":3027,"_3028":152,"_20":-5,"_3029":-5,"_3030":3031,"_3032":146,"_3033":-5},[3034],"hasRecentReleaseNote","membership","referenceId","01a13f50-1e98-760d-bb4c-8201664cc3b8","unreadReleaseNotesCount","userAccount","bot"]

Discussions

Glama MCP Gateway

Resources