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

Adds real behavior beyond the annotations: progress renews the lease, question terminates the run and triggers a follow-up from the delegating AI, and a rejected report means stop. It also implicitly reveals that the response carries a 'continue' flag and follow-up messages, which matters with no output schema. Gaps remain — nothing explains what 'stop' entails concretely, what a rejection payload looks like, or the retry/idempotency contract that idempotentHint=true implies (only the schema's report_id note covers that). No contradiction with annotations: the tool is a state-changing write (readOnlyHint=false) and the description treats it as such.

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?

A single sentence with zero filler, and the most operationally urgent fact (a rejected report means stop; stop when continue is false) is given prominence. The parenthetical enumeration of kind values is dense enough that the first read can be misparsed as prose rather than a list of allowed values, which costs it the top score.

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 12-parameter tool with nested objects and no output schema, the description covers the request side well but says almost nothing about what comes back — the continue flag, the seq/follow-up cursor used by after_seq, or what a rejection looks like — all of which the agent must react to correctly. 'Stop' is the terminal rule yet is left undefined. Adequate to submit a well-formed report; not complete enough to handle the loop it drives.

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 only 33%, but the description compensates by mapping kind values to their payload fields: progress→line, question→question, result→summary and evidence, failed→reason. That conditional field guidance is exactly what the JSON schema cannot express and is the highest-value content here. It adds nothing about lease_token, task_id, after_seq, assistant, or the activity object, which the schema documents on its own.

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 enumerates the four report modes (progress, question, result, failed) with a concrete effect for each (lease renewal, ending the run, summary+evidence, failure reason), which tells an agent precisely what this tool submits. It never states the verb+resource outright ('submit a report on your claimed task') — that framing comes from the title — and it does not distinguish itself from task_status/task_result/task_review, so it stops short of 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?

Gives per-mode usage conditions and hard stop rules: 'stop at once when continue is false', 'ends this run', and 'Any rejected report means stop'. What is missing is routing against alternatives — nothing says when to use this versus task_result/task_review or when a worker should call worker_claim again.

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