LLM Audit

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

Beyond the readOnly/idempotent annotations, the description discloses important behavior: it can return a 'running' status, it requires polling, and it details what the ready result contains (verdict, won counts, competitors, providers). This gives the agent a reliable model of the tool's dynamic behavior.

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?

Three sentences with no wasted words: purpose, polling behavior, and ready-state output are each addressed succinctly. The most important usage constraint is front-loaded.

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 single-parameter polling tool with rich annotations, the description covers the full call cycle: when to call, how long to wait, what status to expect, and what the final result contains. No output schema exists, but the description adequately compensates by listing the result components.

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 input schema already documents runId with 100% coverage ('From start_visibility_check'), and the description reinforces the relationship without adding deep parameter-level detail. Baseline 3 is appropriate because the schema carries the parameter semantics.

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 clearly identifies the action as collecting the result of a measurement started with start_visibility_check, and it naturally distinguishes this tool from its sibling, which initiates the measurement. The verb 'collect' plus the specific resource make the purpose unambiguous.

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 description provides concrete usage context: it must be called after start_visibility_check, and it should be polled every 20 to 30 seconds while status is running. It does not explicitly list exclusions or alternatives, but the single sibling relationship is clear enough.

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

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

The description goes beyond the annotations by disclosing that the tool makes live calls to OpenAI, Anthropic, and Gemini, reports actual answers rather than claimed behavior, is rate-limited per domain, and returns either a runId or a quota_reached signal. It also implies the operation is asynchronous, which is important for invocation behavior.

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, each carrying necessary information: the core measurement, the method and truthfulness, the follow-up flow, and the quota behavior. It is front-loaded with the main purpose and contains no filler.

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?

Given there is no output schema, the description adequately explains the return behavior (runId and quota_reached) and the follow-up call. It covers external side effects, constraints, and async timing, so an agent has enough context to invoke it correctly.

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 input schema already documents all six parameters with descriptions, achieving 100% coverage, so the description doesn't need to repeat them. The description adds no parameter-specific guidance beyond the general mention of 'domain' in the quota rule, so the baseline of 3 is appropriate.

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 clear, specific objective: measure whether a brand appears in AI assistant recommendations by asking real buyer questions to OpenAI, Anthropic, and Gemini. It distinguishes itself from the sibling get_visibility_check by noting this tool starts a run and returns a runId, while the sibling retrieves the result.

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 provides concrete usage guidance: call this to start a measurement, then call get_visibility_check with the runId after about a minute. It also states constraints (free, no signup, one measurement per domain every 30 days) and the quota_reached response, so an agent knows when this tool is or isn't available.

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