code_review_architecture

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

Annotations already declare read-only, idempotent, non-destructive, closed-world, so the safety profile needs no restatement. The description adds the traits annotations cannot capture: results are heuristic (not authoritative), scope is pasted text only, and the output is advisory rather than a scanner substitute. It stops short of saying what a finding looks like or that very large files are capped, but with an output schema present that omission is minor.

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 sentences, front-loaded with the operation and scope, followed by a dense but readable checklist. The one-word qualifier 'Heuristic' in the lead position correctly sets expectations early. Slight cost: the inline enumeration reads as a run-on, though no sentence is padding.

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 an output schema that explains the return value and annotations that cover safety, the description supplies the missing essentials: input must be pasted text, any language is accepted, and findings are heuristic. The remaining gap is the language/filename parameters and the input size limit, which an agent with a large file could get wrong.

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%, so the description must carry the load, and it only partially does: 'pasted as text' clarifies that code is a literal string body (not a path or URL) and 'any language' gestures at the language enum. Neither filename nor the auto-detection behavior of language='auto' is mentioned, and the 60,000-character cap goes unstated.

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?

Specific verb (scan) plus resource (source code pasted as text), with an enumerated checklist of what is detected (secrets, insecure calls, concatenated SQL, disabled TLS, hardcoded endpoints, timeouts, retry hazards). No sibling covers pasted-code security scanning — the closest ones (dependency_manifest_review, cloudformation_review, mcp_config_review) target other artifacts, so the tool is distinguishable without opening 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?

'Not a linter or SAST replacement' is an explicit exclusion that tells the agent when the result is insufficient. However, no positive trigger is given (e.g., use this before cloudformation_review / iam_policy_review, or when no repo access is available), and no alternatives are named by name.

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