Decision Integrity

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

Annotations declare readOnlyHint=false, openWorldHint=true, idempotentHint=false and destructiveHint=false, so the safety profile is partly covered. The description adds one genuinely useful behavioral fact — 'Exactly 1 USDC per call' — plus the scope of the audit. However, with no output schema it fails to say what the evaluation returns, what happens on a failing audit, or whether the fee applies regardless of outcome.

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 core capability and the per-call price are stated up front, which is good front-loading, but the string of seven near-duplicate interrogatives is repetitive padding that restates the same evaluative stance in different phrasings. Size is defensible for a complex tool, but the tail sentences do not each earn their place.

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 ten required deeply nested objects, 0% schema field descriptions, and no output schema, the description explains what is examined but not what comes back. An agent cannot learn the shape or meaning of the result (verdict, findings, per-item diagnostics) or whether the USDC charge is conditional. Given the complexity, this is a significant shortfall.

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 reported at 0%, so the description carries the burden of explaining the ten required nested parameters (action, claims, evidence, warrants, authority, qualifiers, unknowns, conditions). It only gestures at the concepts loosely via the failure-mode vocabulary and never maps any guidance to a specific field such as `warrants[].conclusionHash` or `authority[].effect`. It does not compensate for the coverage gap.

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 opening sentence names a specific verb (evaluate) and resource (evidence, warrant, authority for a requested action), and the second sentence enumerates the concrete failure modes it detects (unsupported inference, qualifier stripping, erased unknowns, broken provenance, exceeded decision boundaries). This is far more specific than a restatement of the name. No siblings exist to differentiate from, so the ceiling is 4 rather than 5.

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 closing rhetorical questions ('Does this evidence warrant this conclusion?', 'Is this action authorized?') implicitly signal when to reach for the tool, but no explicit when-to-use/when-not guidance, prerequisites, or alternatives are given. Usage must be inferred from the list of detection targets.

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