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

With no annotations, the description carries the full burden of behavior disclosure. It provides several specific behavioral details: default processing is standard, deferred processing is optional and limited to eligible pins, there is no Flex→Standard fallback, and 429 capacity errors may trigger a retry. It also notes that the operation is a single logical model execution without truncation rewriting. This is substantial transparency for a tool with no structured annotations, though it doesn't address side effects or data safety beyond the retry note.

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 description is compact—three sentences—and front-loads the endpoint and key purpose. It efficiently packs in behavioral details (deferred processing, fallback, retry) without excessive verbosity. The jargon like 'Flex' and 'scientific receipt' might require domain knowledge, but the structure is well organized and avoids redundancy.

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 the tool has an output schema (which clarifies return format), the description covers the main operational aspects: endpoint, processing modes, retry behavior, and fallback guarantees. It does not detail the body structure beyond hints, but since the body is flexible and no parameters are required, the information level is reasonable. The absence of explicit authentication or rate-limit details is mitigated by the retry mention. Overall, it is fairly complete for an inference call.

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 schema has only a free-form 'body' parameter with 0% coverage, so the description must compensate. It does add meaning by mentioning 'exact pin' and the 'processing=deferred' option, which are likely body fields. However, it does not provide a full schema of expected fields, their types, or required vs optional, leaving the agent to infer the structure. The description gives just enough guidance for a basic call but lacks completeness for a free-form object.

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 identifies the exact endpoint (POST /v1/inference/pinned) and the core purpose: running a pinned inference with a 'scientific receipt'. It distinguishes itself from sibling tools like hydracept_run (generic run) and hydracept_pinned_get (retrieve) by emphasizing 'exact pin' and 'no truncation rewrite', though it doesn't explicitly name those siblings. The verb+resource is clear enough for an agent to grasp the operation.

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 implies usage for pinned inference runs and mentions an optional processing mode (deferred), but it does not explicitly state when to choose this tool over alternatives such as hydracept_run or hydracept_pinned_bulk. It lacks a clear 'use this when X, otherwise use Y' directive. The mention of 'eligible OpenAI Responses pins' gives some context, but the usage boundaries are implicit rather than explicit.

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