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

Annotations already declare readOnlyHint=true and destructiveHint=false, so 'Writes nothing' is largely redundant. The description does add the non-obvious operational constraint that a trusted rendered prompt snapshot must exist, which is behavior beyond the annotations and useful for predicting failure.

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?

Two sentences with the core action front-loaded and the constraint following. No filler, no restating of the title, and every clause carries 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?

There is no output schema, but 'unified diff' adequately signals the return shape, and the snapshot precondition covers the main failure mode. Minor gap: it doesn't say what happens when the precondition fails (error vs. empty diff).

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% and each parameter carries its own description, including the superadmin-only caveat on customerId. The description adds no syntax, format, or edge-case detail about conversationId or proposedPromptContent, so the baseline 3 applies.

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 (preview), resource (unified diff), and the two operands being compared (trusted prompt snapshot vs. proposed prompt content). This clearly separates it from siblings like fix_worst_run or get_run_finding_context, which either apply a fix or fetch context rather than showing a diff.

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 precondition 'only works when the run has a trusted rendered prompt snapshot' tells the agent when the tool is applicable, which is genuine usage guidance. It stops short of naming the sibling to use instead when applying the change (e.g., fix_worst_run) or when no trusted snapshot exists, so it lacks explicit alternatives.

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