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

Goes well beyond the annotations by disclosing the reply shape (brief reply, at most one key question, permission-scoped suggestedReferences), the failure behavior ('顾问失败时如实说明'), and the consent boundary that a recommendation is not an association and approving a draft is not authorizing a write. Slight tension with readOnlyHint=false, but that hint plausibly covers non-idempotent in-session draft editing plus an LLM call, while the description clarifies the persistence boundary rather than denying it.

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?

Dense but front-loaded with the most consequential constraint (no save/publish/send) before the conditional routing rules. Every clause carries a distinct rule; the only cost is run-on phrasing that bundles several conditions into single sentences.

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 complex, nested-schema, open-world advisor with no output schema, the description supplies the missing return contract (reply, one question, suggestedReferences) and the workflow/gating rules the agent needs. Some nested draft-field behavior remains undocumented, but nothing that would cause a plainly wrong call is omitted.

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 carry the semantics, and it does for the fields that drive correct invocation: intent and its ADAPT_PURPOSE value, the relevant context fields, and how suggestedReferences relate to draft.references after confirmation, plus which draft fields (purpose, targetUserId) are retained. It leaves other context keys (needId, productId, focusNeedId) and most draft fields unaddressed, so it compensates substantially but not completely.

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?

States a precise verb and resource: it drafts or revises a cooperation-plan draft from user facts, and explicitly scopes the operation as non-persistent ('不保存、公开或发送'). It also differentiates itself from the persistence sibling ('只有用户明确要求保存才调用保存工具'), so an agent can distinguish this from save_cooperation_plan without opening either schema.

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?

Gives concrete routing rules: pass intent=ADAPT_PURPOSE when working from an old draft or switching purpose, pass context.entryPoint/peerId/conversationId when an object is already selected, and defer to the save tool only on an explicit user save request. When-to-use, when-to-switch, and the alternative tool are all named rather than inferred.

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