"Remote desktop control and assistance tools" matching MCP tools:

• compras_listar_promptsA

  MCP Compras.gov.brA

  Lista os MCP Prompts disponíveis com nome, descrição e argumentos. Tools de descoberta para clientes (como o Claude.ai web) que ainda não expõem UI para prompts. Em Claude Desktop / Cursor / MCP Inspector, prompts aparecem em UI dedicada — esta tool é um caminho alternativo, não substituto. Use depois `compras_obter_prompt(nome, argumentos)` para renderizar um prompt específico. Retorno: { "total": int, "prompts": [ { "nome": str, "descricao": str, "tags": [str, ...], "argumentos": [ {"nome": str, "descricao": str | None, "obrigatorio": bool}, ... ] }, ... ] }

  ConnectorNo auth

• input.click_xyA

  GlasswarpA

  Click at native screen coordinates (0…native_width-1, 0…native_height-1 from screen.observe). Last resort when no suitable UIA target exists — prefer input.click_target. Never use JPEG/downscaled pixel coords. Side effect: real mouse click on the remote desktop.

  Connector

  Destructive

  No auth

• input.send_actionsA

  GlasswarpA

  PREFERRED multi-step tool: run 1–10 predictable UI actions in one call (input.click_target, input.click_xy, input.type_text, input.send_keys, input.drag, input.scroll). Side effects: all actions execute on the remote desktop; fails fast before sending if any action is invalid. observe_after defaults true (verification observe: text+targets; set observe_image=true for JPEG). Do not batch across unpredictable waits (page loads, installers, modals) — single-step those. Prefer this over chaining solo click/type/keys tools.

  Connector

  Destructive

  No auth

• list_componentsA

  PCFHubA

  Browse the PCFHub catalog by category, tag, author, control type, host or text, paged and sorted. Use it to enumerate or narrow the catalog; use search for a plain-language question about what a control should do, and get_component for one control in full. Filters take slugs, which list_taxonomy supplies. Nothing here fails on bad input: an unrecognised sort falls back to the default, and filters matching nothing return an empty page — so read `total` and `sort` from the result rather than assuming the filters applied.

  ConnectorNo auth

• export_ticket_to_providerA

  Meta CouncilB

  Export the exact previewed owned ticket to Jira or Linear, using expected_export_hash from preview_ticket_export and the same destination. Requires explicit integrations:write, tickets:write, and independent human approval of the complete preview. Persists the operation before possible remote writes. After an uncertain response, inspect list_ticket_provider_effects and reconcile the original intent; never blindly export again. Remote confirmation is observed, not atomic protection against concurrent remote writers.

  ConnectorNo auth

• federated_assetA

  trip2gB

  Get one attachment of a remote note — a file, not text. Canonical loop: federated_note_html(kb_id, path) -> for each name in its assets: federated_asset(kb_id, note_path=<that path>, name=<name>). Returns the asset's size and sha256 and a short-lived download URL on this server that streams the file from the remote base; fetch it with curl (the URL needs no other credential) and check the sha256. The file itself never comes back in the tool result. Only a directly connected base: kb_id without '/'.

  ConnectorNo auth

• start_bridgeA

  Blender BridgeA

  Start the authenticated Desktop-to-Blender Bridge service on the receiving computer. Requires existing account access and may start Desktop first. Reuses a healthy Bridge; recovery can stop and replace an unhealthy owned Bridge process, interrupting its connections. Returns the control action's status. Use when the user needs the Bridge running; get_app_state or check_connection inspects status without starting it, stop_bridge stops it, and open_blender separately launches Blender.

  Connector

  Destructive

  OAuth

• set_task_fieldsA

  CheckFlowA

  Answer one or more of a task's form fields. Send every answer in one call rather than calling once per field: writing them together costs one read, one round of binding propagation and one notification however many fields moved. Every value is checked before any of them is written, so a batch either applies completely or changes nothing. Naming the same field twice is refused rather than applied in order. The reply is the task's whole field set, not the fields that were written, because a write propagates: a Data Set-linked control fills in every other control bound to the same record, and a dynamic value copies into the controls pointing at it. Refused for a field a condition is hiding, for a control that only displays something (Text, Image, Video, Embed, Separator, SendEmail, LinkedChecklist), on a task behind an incomplete halt task or assigned exclusively to other people, and for controls that hold no single value -- tables and file uploads have tools of their own. A field that reads from a Data Set is answered with a record of its view, by recordKey or by the words the option shows. At most 50 fields in one call.

  ConnectorAPI key

• queue_checkoutA

  SorsaB

  Queue an Amazon checkout on the user's own desktop Chrome (requires the Sorsa extension installed and signed in there). Call when the user says e.g. 'buy 3 of this' or 'replen this ASIN at up to £20'. mode 'review' (default) drives to Amazon's review page and STOPS — the user clicks Place Order themselves. mode 'auto' places the order, but only if the user has enabled Auto checkout in the extension popup AND the order total is within max_price × qty (max_price is required for auto). max_price is ALWAYS in GBP (£) per unit, whatever the marketplace — the server converts it to the local currency at the current rate, so '£20 max on amazon.de' just works. pace 'calm' (default) clicks at a relaxed human pace — use 'fast' only for time-sensitive deals. domain is the marketplace: GB, US, DE, FR, ES or IT. The job expires after 15 minutes if the user's desktop isn't online to claim it. Always tell the user what was queued and that their desktop Chrome will open it.

  ConnectorNo auth

• tools.find_for_taskA

  Revuo.ai MCPA

  Find the best MCP tools for a given task. Describe what you want to accomplish (e.g. 'manage github issues', 'send slack messages', 'query a database') and get ranked results — each result is one (product, tool) pair, since same-named tools across providers are NOT interchangeable. By default only shows tools available via remote (network-hosted) servers. Audience: agent builders looking for installable MCP tools — for B2B SaaS recommendations, prefer products.search or products.find_by_capability. Ranking basis: relevance score desc (+1 per term hit, +3 for full-phrase hit); within an equal relevance tier, callable-now (Open access) + remote + healthy tools are boosted ahead — never a hard filter (a product-bound tool stays in its own money query), then productName. Paid tier is NOT a ranking input — it appears only as an annotation. Every result carries { position (1-based), rank (0..1; 1.0 = top, scales linearly down by ordinal position) } plus preflight annotations { accessModel (open|keyed|account|commercial), healthScore (0-100, nullable), readOnly, destructive, callableNow (true = usable anonymously right now), setup (one-line 'how to get access' pointer for non-Open tools) }. Response: { task, matchType (none|exactSlug|canonicalSlug|nlpFallback), resolvedCapabilities[], capabilityProducts[] (B2B SaaS products that claim the same capability — empty when matchType=none), results[] (MCP-tool-level matches), buckets { callableNow, requiresSetup } }. Each result: { position, rank, normalizedName, displayName, description, inputSchema, relevance, accessModel, healthScore, readOnly, destructive, callableNow, setup, productSlug, productName, serverQualifiedName, isRemoteCapable, tier, unverified, verifiedAt }. Empty case: { task, matchType, message, suggestedQueries[], suggestedCapabilities[] }.

  ConnectorNo auth

• tools.find_for_taskA

  RevuoA

  Find the best MCP tools for a given task. Describe what you want to accomplish (e.g. 'manage github issues', 'send slack messages', 'query a database') and get ranked results — each result is one (product, tool) pair, since same-named tools across providers are NOT interchangeable. By default only shows tools available via remote (network-hosted) servers. Audience: agent builders looking for installable MCP tools — for B2B SaaS recommendations, prefer products.search or products.find_by_capability. Ranking basis: relevance score desc (+1 per term hit, +3 for full-phrase hit); within an equal relevance tier, callable-now (Open access) + remote + healthy tools are boosted ahead — never a hard filter (a product-bound tool stays in its own money query), then productName. Paid tier is NOT a ranking input — it appears only as an annotation. Every result carries { position (1-based), rank (0..1; 1.0 = top, scales linearly down by ordinal position) } plus preflight annotations { accessModel (open|keyed|account|commercial), healthScore (0-100, nullable), readOnly, destructive, callableNow (true = usable anonymously right now), setup (one-line 'how to get access' pointer for non-Open tools) }. Response: { task, matchType (none|exactSlug|canonicalSlug|nlpFallback), resolvedCapabilities[], capabilityProducts[] (B2B SaaS products that claim the same capability — empty when matchType=none), results[] (MCP-tool-level matches), buckets { callableNow, requiresSetup } }. Each result: { position, rank, normalizedName, displayName, description, inputSchema, relevance, accessModel, healthScore, readOnly, destructive, callableNow, setup, productSlug, productName, serverQualifiedName, isRemoteCapable, tier, unverified, verifiedAt }. Empty case: { task, matchType, message, suggestedQueries[], suggestedCapabilities[] }.

  ConnectorNo auth

• findagent_reintrospect_mcpA

  FindAgentA

  Publish a NEW version of YOUR OWN published MCP server by RE-INTROSPECTING its remote endpoint. FindAgent re-lists the remote server's current tools (over a sandbox-gated, SSRF-hardened scan), compares them to your live listing, and — if the tool surface CHANGED — submits a new version for admin review with an auto-drafted changelog. Your LIVE listing keeps serving until the new version is approved (never auto-unpublished). If the tools are unchanged it is a no-op. If your server has started REQUIRING AUTHENTICATION, the tools can't be re-listed but the listing's 'needs your own credentials' disclosure is corrected in a new version, with your tool list left unchanged. If the remote can't be reached or the scan is unavailable, nothing changes (reported back). Pass the listing slug (you must own it); optional bump (patch|minor|major, default patch) + changelog override. The server URL is read from your stored listing — nothing executes on FindAgent. Owner-only; for MCP servers only (an agent with code uses the code wizard, an agent built here uses findagent_bump_version).

  ConnectorOAuth

• submit_enquiryA

  siteA

  Submits an enquiry to Pest Control Prices — NOT a purchase, NOT a guaranteed quote. Step 1: call with the answers (keyed by field key from enquiry_fields) and consent=true; it validates and returns a summary, the consent line and a confirmation token — show the person the summary and the consent line. Step 2: only if the person agrees, call again with the same answers, consent=true and the confirmation token; the enquiry is then submitted, and the person receives an email with a link they must click before any provider sees it. Consent means the person has read and agreed to: "Send my contact details to up to three relevant local pest control technicians, chosen by Pest Control Prices, who'll contact me directly. I'll be emailed each firm's name as my details go to it, and can stop any more firms getting them."

  ConnectorNo auth

• register_mcp_serverA

  AnkraB

  Register an external MCP tool server for the organisation. Organisation admins only. Prefer a curated integration: call get_mcp_catalog and pass its adapter_key so the platform's read/write classification of the vendor's tools applies; omit adapter_key for a custom server. Remote URLs must be public http(s) endpoints (the SSRF policy refuses private and link-local addresses). Credential headers ride env as header-name to value entries; a sensitive-looking key (token, api_key, authorization, ...) REFUSES plaintext - the user must store the secret through the secret-slot flow and you pass the returned ${SECRET_SLOT:<id>} sentinel as the value. Never ask the user to paste a secret into the conversation. New servers default to the read_only tier; tools become callable only after grant_mcp_tool.

  ConnectorOAuth

• type_textA

  ReMCPB

  Use this when the task calls for reliable Unicode text entry into a native semantic target or focused native control. Prefer method=auto so Accessibility is used when the target supports verified semantic editing. For Chromium page DOM, use browser_action instead of forcing desktop accessibility text input; auto fails safely when a semantic element rejects the write rather than claiming an unverified paste. Use method=keys only for deliberate physical key-by-key input, and keyboard for shortcuts/navigation/control keys.

  Connector

  Destructive

  OAuth

• read_mediaA

  Organize with AI (OwAI)A

  Read a photo as an inline image (original or adapted to 5 MiB), or a video as an authenticated HTTPS resource link supporting GET, HEAD and byte ranges. Videos are not base64-encoded into MCP. Use rendition=still only for an inline video thumbnail, not a frame at start_ms. For video playback pass the search result's segment_start_ms as start_ms; it is a seek hint, not a clip. The client must fetch video links with its authentication header (Authorization: Bearer for OAuth tokens or OwAI developer keys; legacy x-app-key is also supported) and support the source codec; a returned link alone does not mean playback or analysis succeeded. OneDrive needs no desktop. Local folders and Apple Photos require the owning app online with remote media access enabled; Photos also needs native permission and may first download/export from iCloud on that Mac. Does not launch windows or retain originals on the server.

  ConnectorAPI key

• spending_over_timeA

  USASpending.govB

  Aggregate spending amounts over time, grouped by fiscal year, quarter, or month. Use this to visualize spending trends, identify fiscal-year-end spikes, or compare spending patterns across years. Note: The API returns fiscal_year as a STRING. Cast to int for numeric comparisons. quarter and month are FISCAL periods (month 1 = October, quarter 1 = October-December). Buckets are clipped to the window, so a calendar-2025 window returns a partial fiscal_year '2025' (January- September only) and a partial '2026' (October-December); neither is that whole fiscal year. DoD contract actions are published 90 days late, so when the window reaches into the last 90 days and DoD is in scope the answer carries a data_note: those months are understated, so don't read them as a drop. awarding_agency must be a TOPTIER agency name ('Department of Defense'). Military departments are subtiers: pass awarding_subagency='Department of the Navy'. A non-toptier name in awarding_agency (or a subagency that does not belong to the given toptier) silently returns all-zero aggregates. At least one filter is required (the API rejects empty filter sets with HTTP 400). Typical usage: pass time_period_start + time_period_end. date_type controls the transaction date window. Omitted (the default) or 'action_date', it selects transaction actions inside the window. 'date_signed' selects the associated award's first signing date; 'last_modified_date' selects transaction update dates. 'new_awards_only' retains in-window actions associated with awards first signed in the window. These are transaction obligation totals, not new-award counts: use new_awards_over_time for counts. Award search/count tools use a different default interval-overlap rule. date_type needs time_period_start and/or time_period_end. program_numbers filters assistance awards by assistance listing (CFDA) number, e.g. ['93.778'] for Medicaid; use it with award_type='grants', 'loans', 'direct_payments', or 'other'. autocomplete_cfda() finds numbers by program name.

  ConnectorNo auth

• request_quota_increaseA

  Control PlaneA

  Ask Control Plane to raise an org quota. Use it after a create or update fails on a quota: the error names it, and a workload's total CPU or memory error means cpu-per-workload or memory-per-workload. File it only when the user wants the increase, and ask for the new max if they did not give one. Per-GVC quotas (workloads-per-gvc, identities-per-gvc, volumesets-per-gvc) need gvc. Every call files a new request, so never resubmit a pending one. Returns the request id and status: approved (the new max applies now; retry the action), pending (Control Plane reviews it, and the user gets an email and a Console notification), or denied (with the reason). Lowering a max, and the number of orgs per billing account, go through Control Plane support.

  ConnectorOAuth

• restart_blender_with_extensionA

  Blender BridgeA

  Close and relaunch the selected receiving Blender with its Blender Bridge extension after the user requests or agrees to a restart. Can interrupt current work; unsaved changes may require confirmation inside Blender. Desktop may start if its control service is unavailable. Returns restart/control status, not a completed scene. Use get_setup_status for passive diagnosis or ensure_blender_extension for installation/repair without requesting a full restart.

  Connector

  Destructive

  OAuth

• get_page_performanceA

  rankparse-mcpA

  Google PageSpeed Insights report for a URL: Lighthouse performance/accessibility/SEO scores and Core Web Vitals (LCP, CLS, INP, FCP, TTFB), with lab and field (CrUX) data. Defaults to mobile; pass strategy=desktop for the desktop profile. Subject to a per-user daily cap (50/day) and a service-wide daily cap; successful responses are cached for 24h.

  ConnectorNo auth

• submit_support_requestA

  Skala — Legal Platform for StartupsA

  Submit a support request to the Skala team on behalf of the user. Call this when the user needs human assistance that AI cannot provide, the question is too complex or high-risk, or the user explicitly asks for human support. IMPORTANT: Always confirm with the user before calling — describe what you will submit and ask for their approval. Before calling, compile the issue from conversation context into the description.

  ConnectorNo auth