ascended-browser

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries the full burden and does: it discloses the two success return shapes (a fresh `page` vs `page_unchanged: true`), ref staleness/verification rules, that sequence stops before dispatch on ambiguity or failed guards, that upload is the only file path and fill rejects file inputs, and what counts as sufficient evidence for a save action. This is unusually complete behavioral disclosure.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

The core purpose is front-loaded, but the body is a dense wall of run-on prose mixing workflow guidance, evidence policy, and parameter semantics with little structure; several clauses (save-evidence proof, private-persistence warnings) could be tightened or moved. Length is partly justified by the large nested schema, but structure is weak.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a tool of this complexity with no output schema and many nested action variants, the description covers return values, verification, guard semantics, batching, uploads, and waiting behavior — everything an agent needs to invoke it correctly.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The description explains the semantics of the hard parameters beyond a bare schema: kind/type alias handling, the exact-ref requirement, option+query for select, checked state, fill_form batching with depends_on, and expect_before/expect_after guards. It does not explain tab_id or the top-level container, and the reported schema coverage is low, so a 4 rather than 5.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

Opens with a specific verb+resource+scope ('Perform one verified action in a leased tab') and enumerates the exact action.kind values, so the agent knows precisely what the tool does. It also differentiates from siblings by referring to page/observe/extract results and browser_extract, so an agent can route correctly without opening other schemas.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives rich when-to-use guidance: select for anything list-picking, check for any checked/selected state, sequence for short reversible multi-step flows (with an explicit 'do not place a consequential final submit in a sequence'), fill_form for stable fields, scroll+duration_seconds for continuous scrolling, until for late effects. It also states when NOT to act (no unrelated controls, no inspecting hidden databases, no shell scripts for stronger proof).

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description says the JavaScript runs 'policy-gated' and 'read-only,' which are meaningful behavioral disclosures: the agent knows it cannot mutate state and that some policy restricts execution. However, it doesn't detail what the policy prohibits, how failures appear, or whether the return is serialized, so some gaps remain for an unannotated tool.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is two dense sentences that front-load the core purpose and then immediately provide alternative routing. It is well-structured and every clause adds routing information, though slightly long for a single-breath read.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 3-parameter, no-output-schema, no-annotation tool, the description covers purpose, constraints, and alternatives well. It does not explain what the function returns or how results are structured, which is acceptable given no output schema, but it also doesn't mention error handling or serialization limits for read-only evaluation.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 67%. The description does not mention the 'arg', 'tab_id', or 'function' parameters, leaving those to the schema. The schema already documents 'arg' and 'function' partially; the missing third parameter description is not compensated in the description. Baseline 3 for moderate coverage with no extra parameter guidance.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

The description delivers a precise verb ('Run policy-gated read-only JavaScript'), the resource ('in a tab'), and an explicit scope constraint ('only when browser_extract or browser_observe cannot return the required fact'). It clearly distinguishes itself from sibling tools by naming them.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides explicit when-to-use (only when other tools fail) and when-not-to-use, with concrete alternative routing: 'use browser_extract instead' for text/labels/controls, and 'browser_extract read=design or read=inspect' for styles/visual properties. No inference is left to the agent.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries the full behavioral burden and does so well: it distinguishes fresh reads from change-only observations, explains continuation semantics (evidence_ref + next_cursor, cursor offset, limit default 8000), notes that selector queries avoid building a snapshot, states invalid selector is an error while no match is an empty result, and discloses redaction of secret values and bounded attribute previews.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with the core purpose, then ordered into scoping, continuation, selector, and mode-specific paragraphs. It is dense and almost every sentence adds operative detail, though the sheer length (especially the read=audit and read=design clauses) makes it slightly sprawling for an agent to scan.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a complex 19-param multi-mode tool with no output schema and no annotations, the description covers the major modes, continuation behavior, and error semantics well. It is not fully complete: the level and navigations parameters remain unexplained, and the default page-read return shape is only partially implied through continuation limits.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 79%, and the description adds substantial meaning for target_ref, selector, attributes, find, from_end, evidence_ref, cursor, instruction, and each read mode. However, params level and navigations have empty schema descriptions and are never addressed in the description, leaving real gaps for a 19-parameter tool.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

The description states a specific verb and resource (extract visible content and structured state from a tab) and explicitly differentiates itself from siblings: 'fresh high-level read, not a change-only observation' versus browser_observe, and it warns not to continue managed-output refs from browser_open/browser_observe/browser_act. An agent can distinguish this tool's role without opening the schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

Extremely explicit routing: it says when to pass target_ref, when to use selector instead of a full read, when to use find, from_end, and each read= mode (console/network/inspect/design/audit). It also gives a clear when-not instruction ('Do not continue a managed-output reference from browser_open, browser_observe, or browser_act; start a fresh extraction').

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries the full burden and does so richly: secrets are never recorded, targets are stored by identity not ref, replayed steps involve no model call, elements are re-found live, and a run halts at the first unverifiable step and reports done/remaining with from_step resume. This is exactly the behavioral context an agent needs before invoking a replay tool.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

The purpose is front-loaded and every sentence contributes information (recording rules, save/run semantics, variable handling, stop/resume behavior, authority constraint). It is dense and fairly long, but there is no filler or repetition.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For an 8-parameter, no-annotation, no-output-schema tool, the description covers the important behavioral ground including failure/resume semantics. The main remaining gap is that it doesn't describe what recorded/list/show return, though those are read-only inspection modes where the schema enum hints at intent.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is already 100%, so the baseline is 3; the description nonetheless adds value by explaining the record->save->run lifecycle, what start_url means for a new page, what variables carry, and that save reports detected variables. It doesn't elaborate on from_step/to_step beyond what the schema already documents.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

The opening sentence names a specific verb+resource ('Record and replay browser work') and the body specifies that recording happens from verified browser_act steps, which cleanly distinguishes it from siblings like browser_act and browser_open. An agent can tell what this tool is for without opening the schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

It gives concrete trigger conditions ('After a task succeeds once, action=save... action=run replays it') and a real-world motivation (next job application, next record in the same form). It also states a usage boundary ('run only a flow whose steps the user's request covers'), though it never explicitly names a sibling alternative for one-off actions.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries the full burden and does well: it discloses that refs are generation-scoped, that they remain valid for browser_act, that auth_state is page evidence rather than a run-stopping condition, and that diagnostics reports errors and failed requests since the previous result. It never explicitly states the operation is read-only or mentions cost/rate characteristics, leaving a small gap.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

The core return shape is front-loaded in the first sentence, followed by behavioral notes and then the routing advice, so an agent gets the essentials immediately. It is dense but every sentence carries information; a few clauses (auth_state, diagnostics) could be tightened without loss.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

There is no output schema, so the description must describe results, and it does: element refs, content_blocks, scroll_containers, auth_state, diagnostics, and the 'elements were not shown' signal. Combined with 80% schema coverage and one enum, the definition is nearly self-sufficient; only explicit read-only framing and pagination/volume expectations are missing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 80%, so the baseline is 3, but the description adds genuine meaning beyond the schema: query searches the whole page (not just the last observation), within scopes to a region, frame or element, and cursor continues the prior ordering. It frames these as one coherent narrowing strategy rather than restating field definitions, though it does not add format/syntax detail for within or cursor values.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

The description states a specific verb and resource ('Observe a live tab') and enumerates exactly what is returned: generation-scoped element refs, on-screen controls first, content_blocks and scroll_containers. That level of specificity separates it cleanly from browser_evaluate, browser_extract and browser_screenshot without opening any schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

It gives explicit routing: 'When the result says elements were not shown, narrow the view instead of reading the DOM with browser_evaluate', and points to query/within/cursor as the narrowing path. It also tells the agent how to treat auth_state. It stops short of contrasting with the other siblings (browser_extract, browser_screenshot, browser_flow), so it is strong but not exhaustive.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries the full behavioral burden and does so: returns once the page has settled, batch results omit page bodies, failed URLs are reported per-row without closing siblings, atomic rollback is opt-in, reuse persists across a resumed task, and attention_required/blocked semantics are spelled out rather than suspending the run.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with the core purpose and dense throughout, with each clause carrying operational detail. It runs long and some sentences (the web_search aside, the blocked/attention handoff) could be tightened, but there is very little filler.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

No output schema exists, so the description must describe returns, and it does: the ordered results list as authoritative for tab_id, the settled page snapshot for single-URL, reused=true signaling, and attention_required/blocked fields. Nothing needed to call or interpret the tool correctly is missing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is already 83%, so the schema documents most parameters. The description still adds real meaning beyond it: reuse defaults and cross-resumed-task behavior, the atomic all-or-nothing tradeoff, the url/urls mutual exclusivity, and the read flag's one-call shortcut. Only 'disposition' is left without prose explanation.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and resource ('open or navigate your current browser tab') plus the secondary effect of waking the workspace. It also implicitly separates itself from siblings by directing content retrieval to browser_extract/observe and discovery to web_search.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives explicit when-to-use rules: one call with urls for several sites, single url otherwise, reuse=false only for a genuinely independent tab, and web_search for broad public discovery. It names the alternative tools and the conditions that select them.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations the description carries the full burden and discloses real behavioral traits: password/payment fields are masked, the result reports the capture viewport size, and the live window may be resized so the reported size should be trusted over earlier observations. It does not cover permissions/ownership rules or how the image is delivered, but the disclosure is well beyond the schema.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loads the core purpose and the 'last resort' caveat before elaborating on parameters, and every sentence carries information. It is dense and a touch sprawling, but nothing is filler.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a screenshot tool with no output schema and no annotations, the description covers the key behaviors an agent needs: masking, returned viewport size, comparison semantics, full_page truncation. Only the delivery format of the image itself is unaddressed.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 83% (baseline 3), and the description adds genuine meaning: 'state captures a ref's :hover/:focus styling', compare_with 'diffs against an earlier picture or another tab and boxes what changed', and the tab:<tab_id> side-by-side example. The 8000px full_page cut is also surfaced. It stops short of explaining ref/label futher than the schema already does.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

States a concrete verb+resource ('Take a picture of an agent-owned tab') and immediately positions it against siblings: browser_observe returns refs, browser_act operates, a picture returns none. An agent can distinguish this from every sibling without opening a schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

Explicitly says it is 'a last resort, not a way to find or operate controls' and names the exact condition for use ('only for what an observation cannot describe: canvas, charts, images, visual layout'), routing the agent to browser_observe/browser_act otherwise. It even adds the build-vs-original workflow via compare_with.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations are provided, so the description carries the full burden. It usefully discloses batch semantics (max 20, per-tab outcomes, one bad id does not strand the rest), but stays silent on whether 'close' is destructive/reversible, how 'sleep' differs in effect, and what permissions are needed.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

Two dense sentences, front-loaded with the three actions, with the batch constraint and failure-isolation behavior stated compactly. No filler.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a mutation-capable tool with no annotations and no output schema, the description covers purpose and batch behavior but omits action-specific requirements (whether 'list' ignores tab_id) and the effect/risk profile of close vs sleep.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is only 33% (tab_id has no description), yet the description compensates by mapping tab_id to one tab and tab_ids to several, and adds the non-schema 20-item cap. It does not clarify why tab_id is required even for 'list', which leaves a small gap.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

States specific verbs (list/close/sleep) against a specific resource (workspace tabs), matching the action enum exactly. An agent can distinguish this from browser_open, browser_act, or browser_observe without opening the schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

Usage is only implied by the action verbs; there is no guidance on when to choose 'sleep' over 'close', nor when to prefer this tool over browser_open/browser_observe for tab state. Adequate but gaps remain for an agent deciding between actions.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries the full burden and does so: it discloses that resizing only affects the real window on Linux and elsewhere the window keeps its size, that a successful set/restore returns a fresh page snapshot, that earlier element positions become stale and new refs must be used, and that restore clears emulation.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

It is front-loaded with the primary purpose and walks through actions in order, but the single long paragraph is dense and the clause about Linux vs elsewhere is easy to miss.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 10-parameter emulation tool with no output schema, the description covers the important behavioral details an agent needs: tab ownership context, per-action intent, platform-specific behavior, snapshot return, and ref staleness. Nothing material is missing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The schema already documents most parameters, but the description adds specifics beyond it: the concrete preset dimensions were listed in the description body, and it clarifies semantics of set vs emulate vs restore. Schema coverage is 70% with 5 enums, so it complements rather than repeats the schema.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

The description names the specific operation (resize the real browser window for responsive-layout checks) and enumerates each action verb with its resource. It differentiates itself from layout-inspection siblings like browser_screenshot and browser_observe by scoping to viewport sizing and emulation.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

It gives clear contextual guidance for each action (set for resizing, restore when the check is done, get for reading). It does not name alternative sibling tools or explain when not to use this one, so it falls short of a 5.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries the full burden and delivers: it polls while auto-verifying, runs the configured solver for user-required challenges, presses the checkbox if needed, and enforces a hard limit ('image puzzles are never solved'). Failure behavior (structured verdict) and side-effect (checkbox press) are disclosed.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loads the core action, then layers in the captcha-before-submit case and the handoff rule. It is dense and long, but each sentence contributes distinct behavioral information rather than restating the name.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

There is no output schema, and the description compensates by describing the return shape ('structured verdict', 'reports whether a token was issued'). Combined with the parameter context this is nearly complete, though it could say a bit more about the verdict's fields.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 75%, with defaults and meanings already documented on max_wait_ms, apply_solver and poll_interval_ms. The description reinforces these by referring to the 'wait budget' and the 'configured solver', adding contextual meaning for how the parameters interact without contradicting the schema.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb+resource ('wait out a bot wall on a tab, then attempt to solve it') and goes further by naming the concrete scenarios (Cloudflare 'Just a moment…', embedded Turnstile/reCAPTCHA/hCaptcha). None of the sibling browser_* tools handle bot walls, so an agent can distinguish it immediately.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives explicit when-to-call conditions: before submitting a form with an embedded captcha widget, and whenever a bot wall may auto-clear. It also states when to stop and hand off to the user vs continue, which is exactly the routing decision an agent needs.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.