Executes a fixed plan returned by preview_changes after user authorization. Verifies session state, runs steps in order, stops on failure, and returns the original result on repeat submission without replaying writes.
执行 preview_changes 返回的固定计划。仅用于用户已授权的操作；先检查会话是否发生变化，顺序执行，失败即停。不是事务；同进程重复提交返回原结果，不重放写入。
| Name | Required | Description | Default |
|---|---|---|---|
| plan_id | Yes |
Changes observed during successful MCP inspections.
v1.0.0Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already flag destructive=true and idempotent=true, but the description goes further: it scopes idempotency to '同进程' repeats returning the original result with no write replay, and discloses that execution is sequential and fail-fast ('失败即停') and is not transactional. That partial-failure/no-rollback semantics is exactly what an agent needs before invoking a destructive tool and is unavailable from the 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?
Four short clauses, zero filler, front-loaded with the action and immediately followed by the authorization constraint and execution semantics. Every clause carries distinct operational information.
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 one-parameter destructive tool with no output schema, it covers authorization, ordering, failure behavior, and idempotency scope. It could still say what an agent receives back on failure or on a duplicate submission beyond 'returns the original result' (e.g., whether partial writes persist), which is the remaining gap.
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?
With a single plan_id parameter and 0% schema description coverage, the description partly compensates by establishing that the id comes from chatgpt_preview_changes, implying provenance and lifetime. It never names plan_id or says whether ids expire or are per-session, so the coverage gap is only partially filled.
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 specific verb+resource (执行 preview_changes 返回的固定计划) and names the sibling that produces the input, so the agent can place it in the preview→execute workflow without opening a schema. It also marks the boundary: it executes a fixed, already-computed plan rather than deciding what to change.
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?
'仅用于用户已授权的操作' is an explicit precondition/exclusion, and '先检查会话是否发生变化' gives a pre-execution step. The dependency on chatgpt_preview_changes is stated, which implicitly routes agents there for planning, though no sibling is named as an alternative for the non-authorized case.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.