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

Beyond the annotations (which declare a non-destructive, idempotent write), the description adds real behavioral context: identity images persist so later generations inherit the face without re-finding the file, and a subject with no images still travels by name and description. It also states a consent/permission gate for real people, which no annotation conveys. Idempotency and the effect of re-calling are left to the annotation.

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?

Core purpose is front-loaded, followed by the when-to-use rule, then the image-reuse behavior, then the constraint — a sensible order with little redundancy. The prose is somewhat literary ('what it lacks as well as what it has', the stitching example) and could be trimmed, but each sentence carries operational content.

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 10-parameter tool with nested objects, the description covers the non-obvious parts an agent would otherwise get wrong: when to propose a subject, the consent limit, how identity references propagate, and the fact that images are optional. Return values are covered by the output schema and most trivial fields by the schema descriptions, so nothing essential is missing.

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 70%, and the description adds genuine meaning to the ambiguous fields: 'description' should record what the subject lacks as well as what it has (with a worked example), and 'identity' accepts either a fresh generation or an upload and becomes the canonical face used in later prompts. Fields such as tags, cover, sheet, avatar and voice are left entirely to their schema text.

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 opening line, 'Name an identity so it can be reused,' states a specific action and the resource's purpose, and the later sentence about faces travelling into tagged generations makes the object concrete. It does not, however, distinguish itself from delete_subject, update_subject, or list_subjects, so the agent gets no explicit boundary against siblings.

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 unusually well signposted: 'Offer one when something the user brought appears in a second generation, never unasked' gives both a trigger condition and an explicit exclusion against proactive creation. What's missing is naming an alternative for the revise case (update_subject) or how this relates to save_subject_performances.

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