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

Annotations only tell us it is non-read-only, non-destructive and closed-world; the description adds material context beyond that: no AI model or token consumption, the original image is preserved, an improved version is created only when useful, existing corrections are reused rather than recomputed, and existing usages are not replaced. What is missing is any statement about permissions, failure modes, or what the returned artifact looks like (no output schema exists to cover that).

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 most important constraint ('UNIQUEMENT à la demande explicite') is correctly front-loaded, but the text is a dense run-on with filler such as 'sans remplacer les usages existants' that adds little actionable meaning. The mix of French prose with English tool names and the trailing clause about modifying a block dilute the structure.

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 mutation tool with no output schema, the description does cover the crucial behavioral facts (non-destructive, creates a new version, no AI cost, reuses prior correction). However, it leaves the two required inputs entirely unexplained and gives no indication of the result handle an agent would need in order to call select_image_version afterwards.

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 description coverage is 0% for two required parameters (assetId, version), and the description never mentions either one. 'version améliorée' refers to the output artifact, not the required 'version' input, so an agent gets no clarification on which version identifier to pass or whether assetId must reference an already-imported asset. With zero schema documentation the description needed to compensate and does not.

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 names a specific verb and resource ('Améliorer une image') and specifies the exact nature of the operation: local correction of brightness, contrast and sharpness, explicitly without AI or tokens. It also distinguishes the outcome from a plain overwrite by saying the original is kept and an improved version is created if useful. It falls short of 5 because the closing clause about 'modifier le bloc voulu' is vague and reads more like a workflow hint than a scope statement.

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?

There is an explicit gate: use ONLY at the user's explicit request, which is a real when-not constraint rather than implied usage. It also routes the follow-up decision to select_image_version for comparing/choosing between versions. The alternative path 'modifier le bloc voulu' is under-specified and does not name an actual sibling tool, leaving one branch ambiguous.

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