review_work

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

Annotations already declare destructiveHint=true and openWorldHint=true; the description adds value beyond them by stating what step=read returns (diff, per-hunk ledger, rubrics to load), that prepare_comments produces a saved draft, and that publishing requires local user approval. It does not mention the all-hunks-accounted-for rejection rule or rate/order constraints, so a 4 rather than 5.

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?

Dense but efficient: nine sentences for a nine-step workflow, with no filler. The opening enum list precedes the statement of purpose, which is slightly less front-loaded than ideal, but nothing is wasted.

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 multi-step tool with a very large discriminated-union schema, no output schema and destructive semantics, the description covers the step space, the return of the read step, the draft/publish gate and the single-review framing. It leaves some inter-step data flow (e.g. how draftId is obtained) implicit, keeping it at 4.

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 one top-level parameter and effectively no top-level description, the description carries the burden and does so well: it enumerates the legal step values and explains each one's behavior (e.g. step=rubric returns full rubric text for clients that cannot load skill files). It adds little about base/mode/target semantics, which the schema itself covers.

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 concrete verb and resource (review changes before pushing or opening an MR/PR) and enumerates the workflow steps, so the agent knows exactly what the tool does. It never differentiates from the siblings review_sonar or plan_ticket_tasks, so it stops 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.

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

It gives an explicit starting point ("Start with step=read") and explains what each subsequent step is for (rubric, checks, blast_radius, prepare_comments, publish), which is clear operational sequencing. It offers no exclusions or sibling alternatives, e.g. when to prefer review_sonar, so it is not a full 5.

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