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

Annotations already declare the tool is read-only, idempotent, non-destructive, and not open-world. The description adds useful behavioral context beyond that: it returns `clear|changed|attention`, at most 3 actions with source and falsifier, and a new cursor, and it advises against sending a manifest. It stops short of explaining rate limits or deeper operational constraints, so 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?

The description is appropriately sized and front-loads the session-start trigger and return shape. However, the telegraphic phrasing (`pick changes`, `source, falsifier`) and compressed fragments make it slightly harder to parse than a cleaner prose version would be.

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 tool with 6 parameters, a nested `policies` object, no output schema, and schema descriptions missing on `policies`, the description does well on input basics and return statuses. It is incomplete because it never explains `policies`, `maxMonthlyUsd`, `requireVerifiedOutcome`, or `stackFingerprint`, which leaves an agent without guidance on those optional but potentially important inputs.

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 83%, so the schema already documents most parameters, including `cursor`, `models`, `packages`, `providers`, and `stackFingerprint`. The description reinforces the cursor and provider/model/package submission pattern but does not explain the nested `policies` object or the `stackFingerprint` parameter, leaving a gap despite the high baseline coverage.

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 session-start purpose: reporting what changed for the caller's stack since the last call, with concrete examples like retired models, price changes, and deprecated packages. It clearly distinguishes this from the sibling tools by routing new choices to `decide` and live errors to `rescue`.

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 explicitly says to use this at session start if the code calls third-party APIs, and it names alternatives for two important cases: `decide` for a new choice and `rescue` for a live error. It also gives submission guidance by saying to send provider, model, and package ids, never a manifest.

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