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

Annotations already declare readOnlyHint=false, destructiveHint=false and idempotentHint=true; the description goes beyond them by explaining why repeats are safe ('repeated draft requests never repeat an ambiguous model call'), that drafts run through a bounded AI budget path, and that the tool will not trigger spend or publication. Still absent are auth/permission requirements, whether `save` overwrites prior revisions, and what the response contains.

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?

Sentence count is low and the safety constraint is front-loaded, so nothing is gratuitous, but the opening sentence is an overloaded three-clause list of actions that reads as a wall of text. Given nine discrete actions, a compact enumerated mapping would have been both shorter and clearer.

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 of this complexity the description does cover the action surface, the negative boundaries and the human-review gate, which is the essential scaffolding. It is not complete: action-selection conditions, concurrency semantics implied by expectedRevision, and return behavior (no output schema exists) are all left to inference.

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% and the single `command` parameter fans out into nine shapes with dozens of nested fields, yet the description never explains a single field: requestKey, expectedRevision, sourceRefs, assets.rights.status/expiresAt, embargoUntil and the caseId pattern are all undocumented anywhere. The prose hints at which behaviors exist but adds no meaning an agent could use to populate the payload correctly.

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 names concrete operations over a concrete resource: saving a source-bound internal press brief/draft, generating editorial language drafts, preparing OpenPR packets and pitch previews, collecting coverage references, verifying source text, and scanning owner-enabled monitoring. That is specific enough to separate it from read-only siblings like get_press_release or get_press_workspace. It loses a point because the operations are packed into two run-on sentences rather than enumerated against the nine `action` constants, so the agent must map prose clauses back to const values itself.

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?

There is strong explicit when-not guidance: no human approval, publication, send, new provider or spend authorization, and publication/send must go through human review at /app/demand/press. What is missing is positive routing between the internal actions (e.g. when to use save vs draft vs prepare_outreach) and any statement of prerequisites such as case ownership or an existing workspace.

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