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

Annotations already cover readOnly, idempotent and non-destructive, so the bar is lower; the description still adds the substantive caveat 'This tool never authorizes spend,' which matters because a retryPrice parameter is accepted. It also discloses the decision taxonomy rather than returning opaque data. It does not say how historyComplete affects the decision, which is the main remaining gap.

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 sentences, front-loaded with the trigger condition, then input requirements, then return values, then the spend caveat. Dense and near-waste-free; slightly compressed wording in the middle sentence ('grounded in available assets or a specific untried change') takes a moment to parse.

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 nested 8-parameter decision engine with no output schema, the description does supply the return taxonomy and the spend boundary, which is above average. It falls short on the meaning and effect of several required/optional fields, notably historyComplete and the price/time-window data, which an agent must guess at.

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 8 parameters including nested objects, so the description must carry weight; it partially does by defining the shape of observations, chronological ordering of previousAttempts, and what counts as a valid candidate fix ('grounded in available assets or a specific untried change'). But it says nothing about take, brief, model, promptUsed, retryPrice, or historyComplete, leaving a substantial share of the inputs undocumented.

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 decision function ('Decide what to do before another render') and enumerates its five possible outcomes (keep, edit first, propose one retry, change approach, review), so an agent knows exactly what it gets back. It does not explicitly differentiate itself from the sibling review_completed_render, which keeps it short of a 5.

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 clear precondition ('Use after inspecting a completed render') and input-shaping guidance (one observation per requirement, chronological attempts, grounded fixes). However, it never mentions the sibling review_completed_render or when to prefer it, and gives no exclusion criteria, so the routing gap is real.

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