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

Annotations declare destructiveHint=true, idempotentHint=true and openWorldHint=true; the description adds material context beyond that — which operations land in the approval queue (orders, agreements, contact exchange, pricing, rule changes, deletions) and that owner_said is written to an audit log. It does not mention the idempotency key's behavior or what a rejection does downstream, so it falls short of a 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?

Front-loaded with the action, then the trigger condition, then the procedure, then the constraint — a sensible order with no filler. The long em-dash enumeration of example decisions is the only part that costs more space than it returns.

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 destructive, open-world mutation tool the description covers the critical unknowns: the precondition, the mandatory human confirmation, and the audit trail. With annotations already carrying the safety profile and no output schema to explain, the remaining gap — what a reject returns or whether approval is reversible — is minor.

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 50%, so approval_id and decision carry no schema-level descriptions; the description compensates by tying approval_id to the apr_ identifier produced by the awaiting_owner response and explaining that owner_said captures the human's verbatim words for audit. client_request_id semantics are left entirely to the schema, which already documents them well.

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?

Opens with a specific verb pair and resource: "Approve or reject a pending approval (apr_)", and immediately conditions it on "after asking your human." The apr_ prefix and the awaiting_owner trigger make it unmistakable which object this acts on, distinguishing it from siblings like respond or resolve_terms.

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?

States the exact precondition (a prior step returned status=awaiting_owner with an approval and summary), the required human-in-the-loop workflow (show summary, get explicit yes/no, then call), and an explicit prohibition ("Never approve on your own judgement"). The when-to-use path is fully specified rather than implied.

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