delete_document

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

Annotations already declare destructiveHint=true and idempotentHint=true, and the description adds substantial context beyond them: what the cascade actually includes (the full mixed export), what survives (input originals, external copies, backups, provider data), the absence of SSD physical-erasure guarantees, interruption/reconnection behavior, and that only a minimal receipt is retained. This is exactly the extra material an agent needs for a destructive tool.

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?

A single dense paragraph, front-loaded with the action and the preconditions before the caveats. Nearly every clause carries non-redundant information, though the run-on packing of interruption behavior and non-deletion guarantees into one sentence slightly hurts scannability.

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 local cascade delete with an output schema present, the description covers scope, preconditions, refusal conditions, and survivor-set caveats thoroughly, so return values need not be explained. The remaining gap is the undocumented meaning/derivation of document_id and inventory_digest, which the agent must infer.

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% across all three required params. The description compensates partially by stating inventory_digest must be current and by clarifying that acknowledge_delete is an agent-relayed flag rather than standalone authentication, but document_id is never explained and the digest's format/source is left implicit. Partial compensation, not full.

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 ('执行本地级联删除') and immediately scopes it against the sibling preview_document_deletion by requiring that tool's output to have been reviewed first. An agent can distinguish this destructive cascade from preview_document_deletion and redact_document without opening any 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?

It states explicit preconditions (user must have requested deletion AND seen the full preview_document_deletion impact), explains that acknowledge_delete is agent-relayed rather than an independent authorization gate, and gives explicit refuse conditions when other sessions or remote executions exist. When-to-use and when-not-to-use are both covered.

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