get_evidence_bundle

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

Annotations already establish read-only, non-destructive, closed-world behavior, and the description adds genuinely new behavioral context: character-bounded output, no paid inference, a token budget, date-fallback ordering with undated sources excluded, and an explicit prompt-injection warning that source text is untrusted data. It stops short of describing the response envelope or ranking detail, so it is strong but not exhaustive.

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 purpose is front-loaded in the first clause and each subsequent sentence adds a distinct constraint rather than restating. The telegraphic style ('Uses BM25/cards and optionally enabled local embeddings, no paid inference') is dense but every clause carries information; minor trimming could improve readability.

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 an eight-parameter retrieval tool with no output schema and no annotation detail beyond the safety hints, the description covers the important dimensions: what is returned, how citations work, how filters and dates are applied, budget limits, and the trust boundary. The remaining gaps are source_id and mode value semantics.

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 0%, so the description carries the full burden, and it meaningfully explains project, date_from, date_to, record_id, and budget_chars ('bounded JSON characters'). However, source_id is never mentioned and mode is only obliquely referenced via 'Uses BM25/cards and optionally enabled local embeddings' with no value semantics, so two of eight parameters remain semantically dark.

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 a specific verb and resource ('Read exact current source quotes with stable citation IDs') and adds the distinguishing property that makes it different from its siblings: verbatim quotable evidence rather than memory summaries or records. It never names an alternative tool explicitly, which keeps it just short of a 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?

It gives conditional guidance for its own parameters ('record_id selects only that current permitted record without ranking or embedding') and states a downstream usage rule ('Write the answer yourself and cite citation_id'), which implies this tool does not synthesize answers. It never says when to prefer it over search_memory or get_record, so the when/when-not comparison is left to inference.

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