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

With no annotations provided, the description carries the full burden and does so: cost (3 credits, refunded on failed reads), error semantics (422 costs nothing, idempotency_key_reused), latency (60-100s, dated measurement), and an explicit trust guarantee ('counted in code, never by the model'). It also discloses output structure (clusters, quotes with URLs, videos read, basis), which is exactly the behavioral context an agent needs before committing credits.

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 purpose is front-loaded in the first clause and the operational facts (cost, errors, latency, read-back) are packed into short dash-separated sentences. Some space is spent restating return-value fields that an output schema already covers, which is slightly redundant, but the interpretive glosses ('never by the model', 'views over the channel's usual') justify most of it.

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 5-parameter, zero-required, credit-charging tool with no annotations, the description covers selection rules, cost and refund policy, failure modes, latency, output semantics, and how to retrieve results. An agent has everything needed to decide to call it and to call it correctly; the existing output schema covers the return shape.

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 already 80%, but the description adds real meaning: it explains that 'subject' resolves to that subject's outliers, that 'channel' means uploads beating the channel's usual views, and that video_ids are read directly. It also interprets output terms like 'multiple' (views over the channel's usual) and 'comments_off'. The api_key parameter is left entirely unaddressed.

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 subject ('the top comments of its outliers ... grouped into questions, pain points and requests') and enumerates the three selector modes (subject outliers, video_ids, channel uploads). It is clear what the tool returns, though it is phrased as a noun phrase describing output rather than an explicit verb+resource statement, and it does not explicitly distinguish itself from siblings like channel_outliers or random_outliers.

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?

'Send this, video_ids or channel' plus 'a request naming nothing or two things is 422' gives an explicit selection rule and failure condition. It also maps to POST /v1/comments/insights and points to get_call for reading results back. It stops short of naming sibling alternatives an agent should prefer for simpler outlier queries.

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