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

Annotations already declare readOnlyHint=false and destructiveHint=true, so the write/destructive nature is covered structurally. The description adds genuinely new behavioral context by stating that the submission version and the grading scale are verified ('Version de la remise et barème vérifiés'), implying an optimistic-concurrency check and rubric validation. It stops short of saying what happens on a version mismatch, whether grading overwrites prior feedback, or whether the change is reversible.

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?

Two short sentences, front-loaded with the action and followed by the validation caveat; nothing is padded. The only weakness is that the compression is so aggressive that the second sentence reads as a fragment rather than an explanation of the verification behavior.

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 five-parameter destructive mutation with no output schema, the definition covers the essential decision (grade vs. request changes) and flags the version/rubric checks, which is the core an agent needs before calling it. It remains thin on failure semantics and on the status-to-score dependency, and the French-only description is inconsistent with the otherwise English tool surface.

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%, so the description must carry the parameter burden, and it only partially does. 'Version de la remise' loosely points at expectedVersion and 'barème' points at score, but id, feedback, and the meaning of the two enum values (graded vs. changes_requested) receive no explanation, and it is never stated that score is optional or whether it is required when status is 'graded'.

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 two concrete actions — grading a submission ('Corriger un TP') or requesting a resubmission ('demander une nouvelle remise') — which maps cleanly onto the status enum. It is more specific than the generic tool name, but it never names the object being acted on (an assignment/submission record) or distinguishes itself from nearby mutation siblings like update_record or update_form_response.

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 two operating modes imply when each path applies, and a reader can infer that 'graded' means grading while 'changes_requested' means asking for a resubmission. However, there is no explicit when-to-use vs. when-not guidance, no prerequisite statement (e.g., that the assignment must already exist in a submittable state), and no alternative tool is named among the very large sibling set.

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