Recall: Search

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

Annotations already establish the safe read profile, but the description goes well beyond them: it discloses the record-to-raw fallback, the citation fields every result carries, the meaning of the `owner` flag by speaker type, and the `status`/`status_basis` provenance model. The explicit warning not to present assistant, agent_report, or audit records as the owner's decision is genuine behavioral context an agent could not infer from schema or annotations.

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 opening sentence front-loads the purpose, followed by the raw-mode hint, the citation shape, and the provenance caveat in a logical order. It is dense and heavily semicolon-chained, which hurts scannability, but each clause carries distinct information rather than restating the name or schema.

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?

With no output schema, the description takes on the burden of explaining results and does so: citation fields, the owner flag, and the status/status_basis interpretation rule. Combined with schema-documented scoping (project, all_projects, limit, type), an agent has everything needed to call it and to interpret the response safely.

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 100%, so the schema already carries the baseline. The description still adds interpretive value beyond the schema: it explains that `raw` bypasses records to hit turns, and that `status` values must be read together with `status_basis` rather than treated as the owner's verdict. The `owner` speaker enumeration (user_typed, user_answer, user_plan_verdict) further disambiguates what a matched result means.

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 ('Search the owner's past AI conversations') and scopes the content it targets (decisions, plans, ideas, tables, findings), plus a documented fallback to raw turns. It is unambiguous what the tool returns at a high level. It stops short of naming any sibling (e.g. recall_decisions or recall_ask) to differentiate the narrower query tools from this one.

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 only routing guidance is parameter-level ('Use `raw=true` to search the turns directly'), and there is no statement of when to prefer this over recall_ask, recall_decisions, or recall_related. Usage is implied by the described scope rather than explicitly bounded. It also gives no prerequisites or exclusions for the project-scoping modes.

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