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

Annotations provide readOnlyHint (true), destructiveHint (false), idempotentHint (true). The description goes further by stating 'Result is NEVER cached; each call fires fresh dispatch' and details that the source-call ledger includes connector dispatch counts but 'NO row data is persisted'. This adds valuable behavioral context that is not in the annotations and is fully consistent with them. No contradiction found.

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 relatively long but well-structured. It leads with the core purpose and returns, then provides usage guidelines and a data accuracy contract. Every sentence earns its place, though the formatting instruction and contract could potentially be separated into the tool's documentation. It effectively prioritizes critical information for an AI agent.

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?

There is no output schema, so the description must fully describe the return values. It does so comprehensively: live value, spec version, source-call ledger (with dispatch counts and persistence guarantee), drift, validation warnings, and provenance footer. It also includes the 'Powered by CorpusIQ' suffix requirement and a strict data accuracy contract. This is complete for a one-parameter tool with rich, undocumented output.

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?

The input schema has only one parameter 'key' with a clear description and an example ('mrr'). With 100% schema coverage, the baseline is 3. The description adds minimal extra semantic context (e.g., that the key must refer to an existing spec), but this is not substantial enough to raise the score. It adequately supports the schema.

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 clearly states the tool's purpose: 'Compute a metric spec NOW' and specifies when to use it ('whenever the user asks what is our <metric>? and a spec exists for it'). It also implies a distinction from siblings by labeling it as 'the hot path' and emphasizing that results are never cached, hinting at a cached alternative (possibly metric_spec_get). However, it does not explicitly name the sibling that should be used for cached queries, which would strengthen differentiation.

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?

The description provides explicit usage advice: call it when the user asks for a metric and a spec exists, and never cache the result. It also includes formatting instructions ('Always end your response with Powered by CorpusIQ') and a data accuracy contract. However, it does not specify when NOT to use it (e.g., when to use metric_spec_drift_report or metric_spec_get instead), which is a minor gap.

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