Refine Control

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

With no annotations, the description carries the full burden and does so well: it discloses the rejection response shape ({accepted, reason, per_co}), the rejection rule (narrowing protection for a mapped objective), that rejection is a decision rather than a transient error that re-wording will not pass, and the background re-judgement of assertions after acceptance including that `superseded_assertions` is always 0.

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?

Purpose is front-loaded in the first line, then modes, then rejection semantics, then the post-acceptance workflow — a logical progression. It is on the long side and the closing note about `superseded_assertions` being always 0 is somewhat tangential, but nearly every sentence carries operational content.

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 high-complexity, AI-gated mutation tool, the description covers mode selection, the rejection contract, the non-retryable nature of rejection, and the required follow-up call to get_sufficiency. An output schema exists, so return-value enumeration is not needed, and the description still names the rejection payload fields.

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 coverage is already 83%, so the schema defines most parameters. The description nonetheless adds real meaning beyond it by explaining the interaction between `description` and `codebase_findings` (each alone triggers a different mode; together they combine), which the schema only hints at with 'optional if codebase_findings provided'.

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 ('Refine a control's description') plus the gating mechanism ('AI-gated CO sufficiency check'), and breaks out the two operating modes by parameter. It is clear what the tool does, though it never names which sibling to use instead (e.g., regenerate_controls, strengthen_controls, remap_control), so differentiation is inferential.

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?

Explicitly tells the agent which parameters select which behavior: provide `description` for a direct proposal, `codebase_findings` for a platform proposal, or both for context-augmented evaluation. It also states the post-acceptance follow-up ('Read get_sufficiency once that re-judgement lands'). It lacks any when-not-to-use guidance relative to sibling tools.

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