learning_rollback

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

The annotations already mark the tool as destructive, and the description adds valuable context by guaranteeing that code, skills, capabilities, and GPT.md are not modified. It also clarifies that the affected state is memory-related, which goes beyond 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?

The description is two sentences, front-loaded with the core action, and the safety caveat is a meaningful addition rather than filler. It is concise without sacrificing the most important behavioral constraint.

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?

The operation is simple and the annotations cover the destructive profile, so the description is mostly sufficient for safe invocation. However, it leaves the source and meaning of candidate_id ambiguous and does not describe return behavior, which is more noticeable because there is no output 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?

The schema defines candidate_id as a required string but provides no explanation, and the description does not map this parameter to a learning proposal or promoted lesson ID. With 0% schema description coverage, the description needed to clarify what candidate_id refers to and whether it distinguishes the two rollback/removal modes.

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 states a specific verb and resource: 'roll back a learning proposal or remove a previously promoted memory lesson.' It also draws a clear boundary by saying Code, Skills, Capabilities, and GPT.md are never modified, which helps distinguish it from tools like self_update or apply_patch.

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 intended use is implied by the two actions (undoing a proposal or removing a promoted lesson), but the description never explicitly explains when to prefer this tool over related siblings like learning_promote or learning_history. It does not name alternatives or state conditions where rollback should not be used.

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