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

Annotations already cover the safety profile (readOnly, idempotent, non-destructive), and the description adds genuinely new behavior: the read is served purely from stored rows so it is cheap to poll, plus a rate-limit backoff rule and a retry-once-then-feedback failure policy. The 'no regulator/model/store call' statement is exactly the kind of side-effect disclosure an agent cannot infer from annotations or schema.

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?

Front-loaded with the core purpose and the cheap-to-poll property, then proceeds through usage, error recovery, and the next step; a concrete example call closes it. It is dense and somewhat run-on in the error-recovery sentence, but every clause carries information an agent needs, so little 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?

With no output schema, the description carries the return-value burden and does so: it says the answer contains the watch's current state plus alerts newer than the cursor, and that each answer returns a next_since value for the next poll. Error semantics, the escalation path, and a pointer to get_tool_guide for the full guide round out what a caller needs.

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 100%, so the schema already documents all three parameters and the baseline is 3. The description adds the round-trip semantics for the cursor ('pass the next_since value each answer returns to read only new alerts') and ties watch_token to its origin in watch_recalls, which clarifies how the two parameters work together rather than in isolation.

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?

States a specific verb and resource ('Reads a recall watch's current state and every alert newer than a given cursor') and immediately bounds the scope of the read ('from stored rows only: no regulator call, no model call, no store call'). This distinguishes it cleanly from siblings like watch_recalls (which creates the watch) and stop_recall_watch (which ends 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?

Explicit operational guidance: use the watch_token returned by watch_recalls, start with since=null, then feed back the next_since value each answer returns. It also routes forward ('Next: stop_recall_watch when the person asks to stop') and gives per-error handling for watch_not_found, rate_limited, and watch_failed, including when to escalate to submit_feedback.

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