search_memory

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

With zero annotations, the description carries the full burden and delivers substantially: hybrid retrieval mechanism, RRF fusion, response item fields, ownership scoping, the INVALID_ARGUMENT error contract for top_k < 1, and client auto-detection from clientInfo. It does not cover every possible trait (auth, rate limits), but the core behavioral contract an agent needs before calling is well disclosed.

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 front-loaded with the core function and every sentence conveys a distinct fact, but the full bilingual duplication doubles the length for any single-language reader. The top_k error contract and client parameter note are appended after the main body rather than integrated with a parameter section, making the structure slightly scattered.

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 4-parameter hybrid-retrieval tool with a bare schema and no annotations, the description covers nearly everything an agent needs: retrieval method, return fields, scope behavior, error contract, and the client shortcut. The remaining gaps — a semantic definition of top_k and a dedicated project explanation — are minor because the parameter names are reasonably self-descriptive and the default of 5 is present in the schema.

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 0%, so the description must compensate. It does well for client (optional, auto-detected from clientInfo), partially for top_k (only the error condition is stated; no explicit 'maximum number of results' definition), and not at all for project, which appears only inside the ownership scoping rule rather than as a parameter definition. This is meaningful but incomplete compensation for a fully undocumented 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 opens with a specific verb+resource pair — '混合检索记忆与知识' / 'Hybrid retrieval over memories and knowledge' — and specifies the mechanism (vector + BM25 keywords, RRF-fused) and result shape (hits with id/content/score/type/source). This clearly distinguishes it from the CRUD memory siblings (write/read/update/delete_memory) by positioning it as cross-source retrieval, and the v2 scope rule further differentiates personal memory from shared knowledge.

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 implies usage through the v2 scoping rule (personal memory only for current client/project, shared knowledge visible to all) and the error contract, but it never explicitly states when to choose search_memory over read_memory or ask_kb. There are no 'use X instead' statements or exclusions. An agent can infer this is the retrieval/search tool among its siblings, but the guidance remains implicit rather than explicit.

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