read_normalized_evidence

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

With NO annotations provided, the description carries the full burden for behavioral disclosure. It conveys that results are 'bounded' (max_items capped at 64 in schema) and that it requires 'exact session binding' (security posture). However, it doesn't disclose read-only safety explicitly, response format/return shape, or whether the read mutates session state. For a read tool with zero annotations, it adds some value but leaves behavioral detail unexplained.

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 an efficient single sentence with no wasted words. Every element ('read', 'bounded', 'normalized synthetic evidence', 'exact session binding') carries meaning. It's slightly abstract in vocabulary but structurally tight. Not worth a 5 because it uses compressed jargon ('normalized synthetic', 'exact session binding') rather than plain language, which slightly hurts legibility.

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?

The tool has a complex single request object with nested schemas (DeclarationRef, SessionIdentityBinding, field enum), an output schema, and no annotations. The description hints at bounded reads and session binding but doesn't explain the request composition, the meaning of 'normalized synthetic evidence', or the return semantics. Given the schema and output schema carry structural detail, a 3 reflects that the description is adequate but leaves meaning-level understanding 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?

The description adds essentially no parameter-level meaning beyond the schema. It mentions 'bounded' (mapping to max_items cap) and 'session binding' (mapping to session_instance_id/identity_binding), but schema description coverage is 0% for the single 'request' argument. The single request object is structurally complex (DeclarationRef, SessionIdentityBinding, field enums), and the description does nothing to explain how to compose it. Baseline 3 for coverage is undermined because zero coverage plus a single complex param shifts burden to the description, which it fails.

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 'Read bounded normalized synthetic evidence with exact session binding' clearly states a verb (read) + resource (normalized synthetic evidence) plus a key qualifier (bounded, exact session binding). It distinguishes from siblings like mutate/revert/prepare, and the session binding element differentiates from read_declaration_draft (which reads the draft, not evidence). Loses a point for not naming the filter/scope mechanism explicitly in prose.

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 states exact session binding as a constraint but provides no explicit guidance on when to use this vs read_declaration_draft, or vs the mutation tools. The sibling list includes closely related read/mutate/validate operations, yet the description offers no alternatives or when/when-not guidance. 'Bounded' and 'exact session binding' imply restrictions but don't instruct the agent on selection strategy.

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