Server Quality Checklist

Tool Scores

• actC
  2.7/5.0
  Behavior3/5

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

  Annotations already communicate readOnlyHint=false, destructiveHint=true, idempotentHint=false, so the description does not contradict them. It adds useful context that actions are bounded and that risky or sensitive actions are blocked, which gives some behavioral color beyond the structured hints. However, it does not disclose what counts as risky/sensitive or what effects occur when an action is blocked.

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

  Conciseness4/5

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

  The description is a single, front-loaded sentence with no filler. It packs the core function and a key constraint compactly. Minor deduction for jargon ('compare-and-act') that could be clearer, but overall it is efficiently written.

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

  Completeness2/5

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

  This is a complex tool with nested objects, five parameters, no output schema, and meaningful annotations. The description does not explain the relationship between target_id, target_epoch, and scope_revision, what compare-and-act means in practice, or what the return/observation behavior is (despite the observe_after parameter). The agent would struggle to invoke it correctly based on the description alone.

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

  Parameters1/5

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

  Schema description coverage is 0%, so the description carries full responsibility for explaining parameters. It mentions none of the five parameters: spec, target_id, target_epoch, scope_revision, or observe_after. The nested spec object with action, wait, value, ref, and timeout_ms is left entirely to field names and enum values. The description adds no semantic weight to the parameter structure.

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

  Purpose4/5

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

  The description states a specific verb-resource pair: 'Perform one bounded compare-and-act browser action'. It also adds a scoping constraint ('bounded', 'risky or sensitive actions are blocked') that helps distinguish it from general browser automation. However, it does not explicitly contrast it with sibling tools like open_click_verify or batch, and terms like 'compare-and-act' are somewhat jargon-heavy.

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

  Usage Guidelines2/5

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

  There is no guidance on when to use this tool versus siblings such as batch, open_url, click_verify, or verify. The phrase 'one bounded' implies it is for a single action, and 'risky or sensitive actions are blocked' hints at limitations, but no explicit conditions, alternatives, or exclusions are provided. This leaves the agent to infer usage context.

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

• capabilitiesC
  2.7/5.0
  Behavior3/5

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

  Annotations already declare readOnlyHint=true, idempotentHint=true, and destructiveHint=false, covering the safety profile. The description adds that the tool returns adapters, policy, limits, and browser mode, which is useful but superficial. It does not disclose whether these values are static, dynamic, or how they might be correlated.

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

  Conciseness4/5

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

  The description is one short sentence with no filler, and the main output ('Return adapters, policy, limits, and browser mode') is front-loaded. It is extremely concise, though this brevity comes at the cost of missing useful detail, which is captured in other dimensions.

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

  Completeness2/5

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

  There is no output schema, yet the description does not explain the structure or format of the returned policy, limits, adapters, or browser mode. The trace parameter is also left unexplained. For an agent to correctly interpret the output, additional context is needed; the current description is incomplete for a tool with no output schema.

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

  Parameters1/5

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

  Schema description coverage is 0% and the description does not mention the 'trace' parameter at all. An agent cannot determine what the trace object is for, when to provide it, or how it affects behavior. With zero schema descriptions and no compensatory explanation, the parameter remains meaningless.

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

  Purpose4/5

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

  The description uses the specific verb 'Return' and lists concrete resources: adapters, policy, limits, and browser mode. This distinguishes it from sibling action tools like open_url or close_target, making its role as a read-only capability query clear. The terms 'policy' and 'limits' remain somewhat ambiguous, but the core purpose is understandable.

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

  Usage Guidelines2/5

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

  The description provides no guidance on when to use this tool versus alternatives, nor any context or exclusions. There is no statement about selecting it when an agent needs capability information or avoiding it for target manipulation, so usage must be inferred from the name and title alone.

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

• verifyC
  2.7/5.0
  Behavior2/5

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

  Annotations already establish that this is read-only, idempotent, and non-destructive, so the safety profile is covered. But the description adds no behavioral detail beyond that, such as whether verification polls, waits, or how failures are reported, leaving a significant gap because there is no output schema.

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

  Conciseness4/5

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

  The description is one short, front-loaded sentence with no filler. The phrasing 'element assertion' is slightly awkward, but the definition is appropriately compact.

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

  Completeness2/5

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

  The schema is structurally rich (three assertion variants) and there is no output schema, so the description should clarify behavior and return semantics. It does neither, and it does not mention the target_id requirement or relationship to sibling verification tools.

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

  Parameters2/5

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

  Schema description coverage is 0%, so the description must compensate for undocumented parameters, but it only summarizes the three assertion categories. It does not explain target_id, or the meaning of fields like operator, state, exact, and ref, which are critical for forming a correct assertion.

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

  Purpose4/5

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

  The description names a specific action (verify) and the resources involved (URL, text, or element assertion), so an agent can tell what the tool operates on. It does not, however, compare this to sibling tools like open_click_verify or wait_for, which weakens differentiation.

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

  Usage Guidelines2/5

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

  There is no guidance on when to use verify versus alternatives such as observe, wait_for, or list_targets. The description gives no context about the intended workflow or exclusion conditions.

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

• batchC
  2.9/5.0
  Behavior4/5

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

  The description adds meaningful behavioral detail beyond the annotations: it stops on failure, never retries a possible write, and scopes execution to one write lease. This complements the destructiveHint=true and idempotentHint=false annotations instead of merely repeating them. It does not fully explain what a write lease entails, but the added failure/retry semantics are valuable.

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

  Conciseness4/5

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

  The description is a single, tightly worded sentence that front-loads the core capability and key constraints. Every clause adds information, with no filler or repetition. It could be more detailed, but as far as conciseness and structure, it is well organized.

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

  Completeness2/5

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

  Given the complex actions schema, four required parameters, and no output schema, the description is too thin. It omits return values, error behavior beyond stopping, lease semantics, and any explanation of target_epoch or scope_revision. This leaves significant gaps for an agent deciding how to call the tool correctly.

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

  Parameters1/5

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

  The schema has 0% description coverage and the tool description does not explain target_id, target_epoch, scope_revision, or the structure of actions. The only parameter-related hint is the 1-20 action count, which is far from sufficient for an agent to construct a valid request. The description fails to compensate for the schema's lack of parameter documentation.

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

  Purpose4/5

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

  The description clearly states the tool executes a batch of 1-20 browser actions under a single write lease, with specific failure semantics. This goes beyond the name and title, conveying scope and constraints. However, it does not explicitly distinguish itself from sibling tools like 'act' or 'open_click_verify', though the batching concept is implicit.

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

  Usage Guidelines2/5

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

  The description implies batching multiple actions together but gives no explicit guidance on when to use this tool versus alternatives like 'act' or 'open_click_verify'. It does not state prerequisites, when not to use it, or how to choose between batch and single-action tools.

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

• captureC
  2.9/5.0
  Behavior3/5

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

  Annotations already declare readOnlyHint=true and destructiveHint=false, so the safety profile is covered. The description adds a genuinely useful trait—'short-lived in-memory'—explaining that the image handle is transient, but it does not go further (e.g., lifetime, retrieval path).

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

  Conciseness5/5

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

  Single sentence, front-loaded with the action verb; no filler or redundancy. It earns its place and is easily parsed.

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

  Completeness2/5

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

  With 5 parameters, no output schema, and zero parameter descriptions, the single sentence is not enough to call the tool correctly. The agent is left guessing the semantics of the three required integer/epoch fields and the handle's lifecycle.

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

  Parameters1/5

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

  Schema description coverage is 0%, and the description does not explain target_id, target_epoch, scope_revision, ref, or return_image. The phrases 'viewport or one element' hint at a target selection but do not map to any parameter, leaving five parameters effectively undocumented.

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

  Purpose4/5

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

  The description states a specific verb ('Capture'), a specific resource ('the viewport or one element'), and an outcome ('a short-lived in-memory image handle'). It is clear and unambiguous, though it does not explicitly contrast with sibling tools such as observe or verify.

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

  Usage Guidelines2/5

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

  No guidance is provided on when to use capture versus alternatives like observe, verify, or list_targets. It only describes what the tool does, not when to choose it or when to avoid it; no exclusions or alternative tool names appear.

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

• extract_tablesC
  2.9/5.0
  Behavior3/5

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

  Annotations already declare readOnlyHint, idempotentHint, and destructiveHint false, covering the safety profile. The description adds 'bounded' and 'safe HTTP(S) links' as behavioral details, which is useful context beyond the annotations. However, it doesn't elaborate on what 'safe' means or how bounding is applied, so it doesn't fully leverage the opportunity.

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

  Conciseness4/5

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

  The description is a single sentence with no filler, front-loaded with the action verb. It's concise and easy to scan. However, it is almost too terse—it omits critical details, so while the structure is good, the content is insufficient, keeping it from a 5.

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

  Completeness2/5

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

  The tool has 5 parameters (1 required), no output schema, and no parameter descriptions. The description is far from complete: it doesn't explain return format, pagination, how matching works, what 'safe links' means, or any constraints. An agent would need to guess or look elsewhere for essential details.

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

  Parameters1/5

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

  With 0% schema description coverage, the description must compensate by explaining the parameters. It provides no information about what 'query', 'max_rows', 'target_id', 'max_tables', or 'max_columns' mean or how they control behavior. The description is entirely silent on parameter semantics, making it impossible for an agent to know how to set them correctly.

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

  Purpose4/5

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

  The description states a clear action ('Return bounded matching table rows and safe HTTP(S) links') and resource (table rows and links), and distinguishes this tool from a full snapshot. However, it doesn't name a specific sibling or explicitly contrast with alternatives like 'capture' or 'observe', so it falls short of a 5.

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

  Usage Guidelines3/5

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

  The phrase 'without a full snapshot' implies a scenario where a lighter extraction is preferred over a complete page capture, but it doesn't explicitly state when to use this tool versus alternatives, nor does it mention any sibling tools. Usage guidance is implied rather than explicit.

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

• wait_forC
  2.9/5.0
  Behavior3/5

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

  Annotations already cover the safety profile (read-only, idempotent, non-destructive), lowering the burden. The description adds that the wait is 'local' and scoped to URL/text/element conditions, but it does not disclose blocking duration, timeout outcome, or behavior when the condition is already satisfied.

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

  Conciseness4/5

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

  The description is a single sentence, front-loaded with the operational verb and a compact list of supported condition types. It contains no filler, though the brevity leaves little room for contextual detail.

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

  Completeness2/5

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

  The condition schema is complex, but the description only abstracts it into three kinds. Required target_id is unexplained, timeout and observe_after semantics are absent, and there is no failure or timeout behavior described; with no output schema, an agent would have to infer key runtime details.

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

  Parameters2/5

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

  Schema description coverage is 0%, so the description must compensate, but it only loosely maps to the condition parameter ('URL, text, or element'). It says nothing about target_id, timeout_ms, or observe_after, leaving key parameters unexplained.

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

  Purpose4/5

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

  The description identifies a specific action ('wait') on a resource ('browser state') and enumerates the condition families it supports: URL, text, and element. It does not explicitly contrast with siblings like observe or verify, but the verb and condition list make the core function reasonably distinct.

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

  Usage Guidelines2/5

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

  No guidance is given about when to use wait_for versus alternatives such as observe or verify. There is no mention of scenarios like waiting after navigation, before assertions, or when a non-blocking check is preferred; usage is only implied by the tool's name.

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

• observeB
  3.1/5.0
  Behavior4/5

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

  Annotations already establish readOnly, idempotent, non-destructive behavior, so the burden is lower. The description adds useful behavioral context: results are 'visible DOM-backed controls', refs are 'strict', and an optional revision-diff mode exists. It does not explain ref stability or dynamic-page caveats, but it goes beyond the annotations.

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

  Conciseness4/5

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

  The description is a single sentence with no filler words and is front-loaded with the primary return value. However, it may be overly terse given the presence of 4 parameters and a dual-mode behavior. It is concise but sacrifices necessary parameter context.

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

  Completeness2/5

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

  Given no output schema and 4 undocumented parameters, the description is incomplete. It fails to clarify what 'strict refs' means, how revision diffs are triggered, what max_depth and max_nodes control, and how the return value should be interpreted. The rich annotations help safety but not operational completeness.

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

  Parameters2/5

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

  Schema description coverage is 0%, so the description must compensate for the 4 undocumented parameters. It does not explain target_id, max_depth, or max_nodes; only 'revision diff' hints at since_revision. This is insufficient for an agent to know how to populate parameters correctly.

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

  Purpose4/5

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

  The description uses a clear verb ('Return') and specifies the resource: visible DOM-backed controls with strict refs, or a revision diff. The title reinforces the domain. However, the 'or a compact revision diff' clause introduces mode ambiguity and does not explicitly distinguish this from sibling tools like 'capture' or 'verify'.

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

  Usage Guidelines2/5

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

  No guidance is given about when to use 'observe' versus alternatives such as 'capture', 'verify', or 'wait_for'. There is no explicit when/when-not language, and the reader must infer from the tool name that it is for inspecting browser state. This is implied usage at best.

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

• open_urlB
  3.2/5.0
  Behavior4/5

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

  Annotations already declare read-only, idempotent, non-destructive behavior. The description adds useful behavioral context: the request is a GET, the URL must be allowlisted, waiting happens locally, and the call can return tables and refs. This goes beyond the structured annotations without contradicting them.

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

  Conciseness5/5

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

  A single sentence front-loads the core operation (GET-open allowlisted URL), then lists the auxiliary behaviors (wait, return tables/refs). No filler or redundancy.

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

  Completeness1/5

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

  For a complex tool with 8 parameters, nested schemas, and no output schema, this description is far too thin. It provides no parameter semantics, no return-value shape, no guidance on wait conditions or extraction limits, and no relationship to sibling tools. An agent would need to inspect the raw JSON schema and infer a lot.

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

  Parameters1/5

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

  With 0% schema description coverage and 8 parameters including complex nested objects (wait, observe, extract_tables, trace, target_id), the description does almost nothing to explain parameter meaning. 'Return tables and refs' hints at extract_tables/observe but leaves wait, trace, timeout_ms, new_target, and target_id entirely unexplained.

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

  Purpose4/5

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

  The description states a specific action — GET-opening an allowlisted URL — and names concrete capabilities: waiting locally and optionally returning tables and refs. It gives a clear resource and outcome, though it stops short of explicitly differentiating from open_click_verify or other opening siblings.

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

  Usage Guidelines3/5

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

  The phrase 'in one call' implies a combined open-wait-extract workflow, giving some sense of when it is appropriate. However, there is no explicit guidance about when to prefer open_url over wait_for, extract_tables, observe, or open_click_verify, nor any exclusion conditions.

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

• list_targetsB
  3.4/5.0
  Behavior3/5

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

  Annotations already cover read-only, idempotent, non-destructive behavior, so the description does not need to restate safety. It adds the notion of 'reconciliation state,' which is not present in the title or annotations, but it never defines what reconciliation state means or how it is represented in the response. This is a moderate gap for a read-only listing tool.

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

  Conciseness5/5

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

  The description is a single concise sentence with a clear subject-verb-object structure and no filler. Every word earns its place, and the most important information ('List controlled browser pages') is front-loaded.

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

  Completeness3/5

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

  For a parameterless tool with strong annotations, the description is mostly sufficient, but 'reconciliation state' is an unexplained domain-specific term that likely matters to the agent. There is no output schema and no hint about the response format, so the description leaves a gap around what the listing actually contains and why reconciliation state matters.

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

  Parameters4/5

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

  The input schema has zero parameters, so there is nothing for the description to explain beyond what the schema already conveys. The baseline for 0-parameter tools is 4, and the description does not introduce any parameter-related confusion.

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

  Purpose4/5

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

  The description states a clear verb ('List') and a specific resource ('controlled browser pages') plus a distinctive output concept ('reconciliation state'). It is clear enough to separate this from acting tools like act, close_target, or open_url, but it does not explicitly differentiate itself from observation-oriented siblings such as observe or capture.

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

  Usage Guidelines2/5

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

  There is no guidance on when to call this tool versus alternatives like observe, capabilities, or verify. The listing behavior is implied by the verb, but the description never states prerequisites, exclusions, or a preferred scenario such as 'use this before selecting a target to close or interact with.'

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

• price_radarA
  3.5/5.0
  Behavior4/5

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

  Annotations already declare read-only, idempotent, and non-destructive behavior; the description adds that the data is cached and comes from a public snapshot feed, implying potential staleness and lack of auth. It slightly oversimplifies the 'Top 5' behavior since max_offers is configurable, but it does not contradict the annotations.

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

  Conciseness5/5

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

  A single sentence with no filler, and the key facts are front-loaded. Every word contributes to understanding the tool's function and source.

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

  Completeness2/5

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

  Despite rich annotations, the description omits essential calling context: how product values are specified, what preset does, what the returned offers look like, and there is no output schema to fill the gap. An agent would need additional knowledge or experimentation to use this tool correctly.

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

  Parameters2/5

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

  With 0% schema description coverage, the description must compensate, but it only hints at max_offers via 'Top 5'. It fails to explain the required product parameter, the preset option, or the trace object, leaving an agent without enough meaning to construct a valid call confidently.

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

  Purpose5/5

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

  The description uses a specific verb ('Return') and names a unique resource: cached Top 5 offers from PriceAI's read-only public snapshot feed. It clearly distinguishes the tool from siblings, none of which mention pricing, caching, or snapshots.

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

  Usage Guidelines2/5

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

  The description gives no guidance on when to prefer this tool over alternatives, nor does it state conditions, prerequisites, or exclusions. An agent cannot tell from the description alone which scenario calls for this tool versus a sibling.

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

• open_click_verifyA
  3.7/5.0
  Behavior3/5

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

  Annotations already declare readOnlyHint=false and destructiveHint=true, so the mutation risk is covered. The description adds the blocking and CDP-disabled constraints, which are useful, but doesn't disclose failure behavior, whether new tabs are allowed (though new_target param exists), or how verification is performed. It doesn't contradict annotations and adds moderate context, warranting a 3.

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

  Conciseness5/5

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

  Two sentences with no filler. The primary action is front-loaded, and the constraints are stated compactly. Every sentence earns its place; no redundant information.

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

  Completeness2/5

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

  Given the complexity (8 params, nested objects, no output schema), the description is far too brief. It doesn't explain the 'ready' condition, how to specify the click target's role and name, how 'expect' works, what 'new_target' does, or the meaning of timeout_ms. An agent would need to infer much from the schema alone, which has no descriptions. This leaves the tool under-documented for correct invocation.

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

  Parameters2/5

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

  Schema coverage is 0%, so the description must explain parameters. It vaguely references click (via 'exactly click one visible button/link/tab') and expect (via 'prove a new URL/text condition'), but doesn't elaborate on the structure of these objects, nor does it mention ready, trace, target_id, new_target, or timeout_ms. For a tool with 8 parameters and nested objects, this is a significant gap.

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

  Purpose5/5

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

  The description states a precise action sequence (open, exactly click one visible button/link/tab, verify a URL/text condition) and distinguishes itself from siblings like open_url (which only opens) and verify (which only verifies). The verbs are specific and the resource is an allowed page, making its purpose unmistakable.

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

  Usage Guidelines4/5

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

  The description explicitly says what actions it blocks (risky, sensitive, form, popup, immediate write-network) and that it's disabled for CDP, giving clear exclusions. It implies it's for a single combined open-click-verify operation but doesn't name alternative sibling tools or when to prefer them, so it's clear but not fully exhaustive.

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

• external_model_planA
  3.8/5.0
  Behavior4/5

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

  Annotations already establish readOnlyHint=true and destructiveHint=false, and the description adds valuable context by explicitly stating the tool never executes and clarifies that it calls an external, pre-configured OpenAI-compatible model. There is no contradiction with the annotations.

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

  Conciseness5/5

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

  The description is extremely compact: one sentence plus a short clarifying fragment. It front-loads the core action and keeps every word informative with no repetition of schema or annotation content.

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

  Completeness3/5

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

  The core purpose and non-execution behavior are clear, and annotations plus the schema's defaults/enums cover safety and allowed values. However, there is no output schema and the description does not explain what the returned proposal contains or how the three parameters shape it, leaving some ambiguity for a 3-parameter tool.

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

  Parameters1/5

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

  Schema description coverage is 0% and the description mentions none of the three parameters (task, page_summary, allowed_tools). The agent is left to infer what inputs to provide from parameter names alone; the description does not compensate for the schema's lack of descriptions.

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

  Purpose5/5

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

  The description opens with the explicit verb 'Ask' and names the resource ('explicitly configured OpenAI-compatible model') and the deliverable ('one schema-validated Fast GUI proposal'). The second clause 'Never executes it' sharply distinguishes this from sibling execution tools like act.

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

  Usage Guidelines4/5

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

  The description makes clear this tool is for obtaining a proposal only and explicitly states it never executes, giving an agent a clear context for choosing it over execution-oriented siblings. It does not name alternative tools or list conditions when not to use it, but the non-execution statement provides strong routing guidance.

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

• close_targetA
  4.4/5.0
  Behavior5/5

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

  Annotations already mark this as destructive, but the description adds meaningful behavioral detail beyond that: only owned targets are eligible, unresolved targets need a special flag, and CDP pages are preserved. These are concrete, non-obvious behaviors that help an agent use the tool safely.

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

  Conciseness5/5

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

  Two sentences with no filler. The core action is front-loaded, and the second sentence packs two important edge cases. Every word earns its place.

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

  Completeness4/5

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

  The description covers the necessary behavioral conditions: ownership, unresolved-target handling, and CDP preservation. However, with no output schema and limited parameter detail, timeout semantics and non-owned target behavior are left unspecified. For a simple three-parameter destructive tool, this is still mostly complete.

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

  Parameters3/5

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

  Schema description coverage is 0%, so the description must compensate. It explains discard_reconciliation behaviorally ('unresolved targets require...'), but it does not clarify timeout_ms or target_id beyond their names and schema constraints. The most important parameter is addressed, but not all.

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

  Purpose5/5

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

  The description opens with a specific verb and resource: 'Close an owned page.' It clearly states the action, the scope of what can be closed ('owned'), and distinguishes this tool from siblings like list_targets and open_url without requiring schema inspection.

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

  Usage Guidelines4/5

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

  The description gives clear context for using the tool and a conditional usage rule: 'Unresolved targets require discard_reconciliation=true.' It does not explicitly name alternatives or state when not to use it, but closing is a unique action among the siblings, so the guidance is adequate though not exhaustive.

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