mcp-review

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

Beyond the annotations, it adds that the tool never edits code, is metered, rejects unchanged/exhausted targets without running judgment, and explains provenance behavior when nothing judged the target. It also reinforces idempotency by telling the agent to reuse client_request_id on retries. This is meaningful behavioral detail with no contradiction.

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 dense but still readable, with four focused sentences. No sentence is wasted; each one adds either a behavioral guard, a retry hint, or a required precondition. It is slightly compressed but appropriate for the tool's complexity.

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?

Given the tool has 5 parameters, low schema coverage, and a nontrivial retry/recheck flow, the description leaves gaps around expected_revision semantics and how the provenance signal should be practically handled. The output schema helps, but the description alone is not fully sufficient for an agent to avoid stale or invalid recheck calls.

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?

With only 40% schema description coverage, the description needed to compensate for undocumented parameters like expected_revision, review_id, and client_request_id. It only hints at client_request_id reuse and does not explain what expected_revision is for or how review_id relates to a prior review.

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 a specific verb and resource: submitting a metered recheck for findings from a completed review. It further distinguishes the tool by stating it is for use after the customer's agent changes the UI, which separates it from the sibling design_review and related review tools.

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 clearly states when to call the tool: after a completed review and UI changes. It also gives exclusions by noting unchanged targets and exhausted recheck loops are rejected, which informs an agent not to use it in those cases. It does not fully name alternative siblings, but the context is clear enough without explicit alternates.

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?

Annotations already declare readOnlyHint=false, openWorldHint=true, idempotentHint=true, and destructiveHint=false. The description adds 'asynchronous, metered,' which is not in the annotations, and reinforces the idempotency via 'Reuse client_request_id on retries.' It also states 'never edits code,' a behavioral guarantee. These add value beyond the annotations. It could mention what happens on failure, but the polling instruction implies a retry pattern. No contradiction with annotations.

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 sentences with no wasted words. The first sentence states purpose and a key behavioral fact (never edits code). The second gives actionable retry and polling guidance. Everything earns its place, and the most important information is front-loaded.

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 submission tool with an output schema (which presumably contains poll_after_ms and review ID), the description covers the essential workflow: submit, retry with same ID, and poll with a specific tool. It doesn't explain how to cancel or recheck, but those are separate tools. The description is complete enough for an agent to successfully initiate a review without missing critical steps.

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 description coverage is 71% (5 of 7 parameters have descriptions). The description itself mentions only client_request_id and references poll_after_ms (an output field). It adds a usage note for client_request_id (reuse on retries), which is already in the schema's description, so the marginal value is low. The other parameters (depth, response_mode, etc.) are adequately described in the schema, so the baseline 3 applies.

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 clearly states the action ('Submit an asynchronous, metered design review') and specifies the resource (tenant-authorized HTTPS preview). It also explicitly says 'This tool never edits code,' which distinguishes it from any code-modifying sibling. The verb 'submit' is unique among the sibling tools (get, recheck, cancel, panel_action), so an agent can unambiguously select it.

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 gives concrete usage guidance: 'Reuse client_request_id on retries, then poll design_review_get no faster than poll_after_ms.' This tells the agent how to handle retries and which tool to use for polling. It doesn't explicitly exclude other siblings (e.g., design_recheck, cancel), but for the core submit action the guidance is sufficient. Slightly stronger would be an explicit 'use this to start a review, use design_review_get to retrieve results.'

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?

Beyond the annotations declaring destructiveness and idempotency, the description discloses best-effort semantics, the terminal-job no-op behavior, and the fact that no customer systems are edited and no review units are consumed. These are material consequences for the agent and are genuinely additive to the structured hints.

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?

Three sentences, nothing wasted. The core action is front-loaded, and the secondary constraints are compactly chained. Every sentence earns its place.

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 two-parameter cancellation tool with an output schema and strong annotations, it captures the essential state constraints, best-effort nature, and safety characteristics. A small mention of confirming the result via design_review_get would have made it fully rounded, but it is not needed to call the tool 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?

Schema description coverage is 0% and the description is silent on both job_id and reason. An agent might infer job_id from the tool name, but the description itself offers no meaning for reason, its optionality, or its use.

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?

Says what it does with a specific verb ('Request best-effort cancellation'), identifies the exact resource (a queued or running review job), and carves out terminal jobs. The backoff semantics and no-edit behavior distinguish it clearly from design_review, design_review_get, and design_review_panel_action.

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?

Provides clear operational context: applicable to queued/running jobs, best-effort, and no-op for terminal jobs. It does not explicitly name sibling alternatives or say 'check design_review_get to confirm a terminal state,' so it stops just short of full when-to-use routing.

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?

The description adds substantial behavioral context beyond the annotations. It explains 'Result reads do not consume review units', which is safety-relevant. It also gives actionable guidance on how to interpret two result fields: 'Act on review.measurements.violations unconditionally' and 'Act on review.findings only when provenance.model_backed is true and coverage.state is full or partial'. These nuances are not present in annotations and are crucial for correct use. There is no contradiction with annotations.

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 compact and front-loaded. It states purpose in the first sentence, then provides specific usage (polling) and a second critical behavioral rule (result units) followed by a third interpretation rule. Every sentence earns its place with no fluff 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?

The description is highly complete for a read-only view tool. It tells the agent about the output views, the impact of read operations on review units, the polling mechanism, and the provenance/coverage gate for findings. Given that the output schema exists and annotations already convey idempotence and safety, nothing essential is missing for correct use.

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 describes 'view' and its enum/default, covering 50% of parameters. The description enhances meaning by connecting the view options to 'compact, focused, or evidence' and specifying that the job_id refers to an 'existing review job'. While it does not provide detailed semantic descriptions for each enum value, it goes beyond the schema's minimal value descriptions and reinforces the job_id context.

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 ('Get'), a clear resource ('design review'), and the exact variants of views available (status, compact, focused, evidence). The phrase 'existing review job' immediately distinguishes it from sibling tools like design_review or design_recheck, which likely create or modify jobs. An agent can reliably identify this as the read/query tool among the sibling set.

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 clearly tells the agent when to use it: to get status or views of an existing review job, and it gives a concrete operational guideline ('Poll no fetched than the returned poll_after_ms'). It does not explicitly state when not to use it or mention alternatives (e.g., 'use design_review to start a new review'), but the context of 'existing review job' implies it is not for creating or canceling. The guidance is clear but weak on exclusion.

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?

The description aligns with the readOnlyHint and destructiveHint annotations by stating it never edits code and reads only. It also discloses specific edge-case behaviors for advisory findings and unjudged reviews, adding detail 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.

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

The description is somewhat repetitive, repeating the read-only idea in multiple phrases ('never edits code', 'Reads only', 'consumes no review units'). It is not excessively long but could be more streamlined.

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?

The description explains return outcomes and read-only behavior, but it lacks details about parameter meaning, error cases, or output structure. Given an output schema exists, the description is adequate but not fully complete.

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 0%, and the description only indirectly references 'finding' without clearly defining job_id, action, or finding_id. The action enum values are not explained, leaving parameter semantics largely unspecified.

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 clearly states the tool routes an interaction from the review panel and returns either a fix or refs to re-verify. It also clarifies the read-only nature, distinguishing it from action-oriented tools.

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 gives some usage context ('interactive review panel') and mentions outcomes like advisory findings, but it does not explicitly state when to choose this tool over siblings such as design_review_get or design_recheck. No direct alternative guidance is provided.

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