search_products【何时用】用户用自然语言找产品/作品时，比如「有没有给独立开发者用的财务工具」「记笔记的极简 app」「Notion 替代品」。返回按相关度排序的产品卡片，含 slug / tagline / 所属主理人。
【只管产品】找「人」（能提供某种价值的主理人）用 search_people；搜需求用 search_needs。
【机制】关键词 + 向量（阿里云百炼 text-embedding-v3）双路并行召回后 RRF 融合，另有 LLM 查询扩展 / 精排，各步可自动降级。结果里的 mode 一般为 hybrid。
【常见 pitfall】问 "什么是独行录"、"如何注册" 这种 meta 问题不要用本工具，那是站点介绍不在数据里。
| Name | Required | Description | Default |
|---|---|---|---|
| q | Yes | 搜索查询，自然语言或关键词 | |
| limit | No | 返回条数，默认 12，最多 30 |
Changes observed during successful MCP inspections. Dates show when Glama detected each change.
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already cover readOnly/idempotent/non-destructive behavior, so the description's burden is lower. It adds useful behavioral context beyond annotations: the hybrid retrieval mechanism ('关键词 + 向量...RRF 融合'), auto-degradation ('各步可自动降级'), and the expected 'mode 一般为 hybrid' field. However, it does not detail error responses or edge cases like empty results.
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 organized into four labeled blocks, front-loading when-to-use and sibling differentiation before mechanism and pitfalls. It is somewhat longer than strictly necessary, but every section earns its place and the structure makes it easy for an agent to scan.
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?
Given there is no output schema, the description compensates by naming return fields (slug/tagline/主理人) and warning against meta queries. The mechanism and sibling routing are covered. A full response shape would be nice, but for a read-only search tool this is sufficiently complete.
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 100%, with both q and limit already documented. The description reinforces the natural-language intent with examples but adds no new parameter-specific details beyond the schema. Baseline 3 is appropriate.
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 clearly states the tool finds products/works ('用户用自然语言找**产品/作品**时') and describes the return output ('返回按相关度排序的产品卡片，含 slug / tagline / 所属主理人'). It also explicitly distinguishes itself from sibling tools, search_people and search_needs, in the '只管产品' section.
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?
Provides explicit when-to-use guidance with concrete natural-language examples, and explicitly routes the agent to alternatives: '找「人」...用 search_people；搜需求用 search_needs'. Additional when-not-to-use guidance comes from the meta-question pitfall ('什么是独行录'、'如何注册'), leaving no ambiguity about scope.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Add one secure layer between your agents and this server.
Each tool has a clearly documented purpose, often with explicit 'when to use' guidance and cross-references, making the vast majority easy to tell apart. A few clusters (get_my_brief, get_my_positioning, get_my_work, get_my_dispatch) and data-overlapping get_my_card vs get_my_profile require careful reading, but descriptions are detailed enough to prevent serious misselection.
The overwhelming majority follow snake_case verb_noun conventions (create_product, update_need, list_my_signups). Minor deviations include noun-only feed names (personalized_feed, random_feed), inconsistency between 'prefs' and 'preferences' in notification tools, and a mix of update_* and set_* for mutations, but the pattern remains predictable overall.
137 tools is an extreme mismatch for any MCP server, far exceeding the 50+ threshold for a score of 1. Even with a broad multi-domain platform, this volume makes tool selection and navigation impractical and heavily burdens the agent's context window.
The surface covers full lifecycles for needs, products, activities/signups, conversations, collaboration goals/tasks, dispatch, profile/onboarding, and supporting resources like companies, parks, policies, and ratings. Deliberate omissions (no user-post creation, no organizer profile editing via agent) are explicitly documented, so core workflows have no obvious dead ends.