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

Annotations already cover the safety profile (read-only, idempotent, non-destructive, closed-world), so the bar is lower, and the description still adds substantive behavior: it reports a verdict plus location per rule, surfaces ungrounded numbers/names/links/quotes as a hallucination signal, and supplies the corrective line for each broken rule. That is more than the annotations convey. It does not discuss cost, determinism, or how many rules it can handle.

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 opening question is well front-loaded, but the core of the description is one enormous run-on sentence with a sprawling parenthetical enumerating two dozen rule types, followed by another tangled clause about unsourced facts. The raw information is relevant, yet it is delivered as a wall of text that is hard to scan, so the structure undercuts the value.

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?

With an output schema, full annotation coverage and fully described parameters, the description only needed to explain what the check examines and when to reach for it, and it does both at length. The main remaining gap is again sibling positioning against fix_bad_answer and grade_prompt, which matters given ten sibling tools.

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 100%, so the schema already documents all four parameters, which sets the baseline at 3. The description tells the agent what the prompt and answer should contain conceptually but adds nothing about the `target` enum, whose purpose ('the model the fixed prompt should be laid out for') is especially unintuitive for a checking tool and is left entirely to the schema.

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 gives a concrete verb and resource: it finds checkable rules in the prompt and reports, per rule, whether the answer kept or broke it, plus flags unsourced numbers, names, links and quotes. That is specific enough to distinguish it from grade_prompt or lint_prompts. It stops short of naming a sibling explicitly, so it does not fully route the agent among the many look-alike tools in this family.

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 gives a use moment ('before an AI answer is used or sent') and a secondary use ('to make a prompt more reliable'), which is real guidance. But there is no when-not guidance and no reference to the obvious alternative, fix_bad_answer, which an agent would need to know whether to check first and fix second, or the reverse.

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