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

Beyond the annotations (readOnly, openWorld, non-idempotent), the description discloses the BYOK auth requirement and key formats, the external provider calls to GPT-4o/Claude/Gemini, the full return shape (per-case verdicts, pass rate, overall PASS/WARNING/FAIL, latency), an adjustable threshold, and the exact semantics of semantic_contains (TF-IDF cosine ≥ 0.4). These are material behavioral facts an agent needs before firing a billable network call.

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?

Front-loaded with the purpose, then inputs, then outputs, then assertion taxonomy and the BYOK caveat — no filler sentences. The em-dash chaining makes it denser than ideal, but 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?

An output schema exists, so return values needn't be restated, yet the description still summarizes verdict structure usefully. Given 8 parameters and external API dependencies, the only meaningful gaps are failure modes (invalid key, provider rate limits, cost per run) and whether the non-idempotent latency stats vary across runs.

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 baseline is 3; the description earns extra credit by explaining the two mutually exclusive image inputs (URL or base64), the question+assertion structure of each test case, and the assertion-type vocabulary with the semantic_contains similarity floor. It doesn't add anything about threshold defaults or the 10-case cap beyond 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?

Opens with a precise verb+resource: 'Run a test suite against a Vision-Language Model', and immediately specifies the inputs (image + N test cases with question/assertion) and the supported providers. This clearly separates it from text-only siblings like prompt_test_suite and run_semantic_tests, though it never names a sibling to route against explicitly.

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?

Usage is only implied: you'd reach for this when you want to evaluate a VLM's answers on an image against assertions. There is no statement of when NOT to use it, no mention of prompt_test_suite / compare_models / run_semantic_tests as alternatives, and no prerequisites beyond the BYOK note.

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