review_practice_attempt

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

Annotations only declare readOnly=false, idempotent=false, destructive=false, openWorld=false, so the description carries real additional weight: it will not rewrite the answer, question, or original grading, refuses to auto re-grade attempts lacking a snapshot, and preserves the old feedback on a revision. The withdraw mechanism and its constraints are also disclosed. This is substantial context beyond the annotations and is fully consistent with them.

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?

Purpose is front-loaded, followed by prerequisites, then per-field semantics and edge cases, so ordering is logical. The prose is dense and run-on in places, and mentions user_opinion_source_revision, which does not appear in the input schema. Every sentence still carries distinct information, so waste is low.

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?

With an output schema present, return values needn't be described, and the description instead covers the workflow, defaults, conflict handling, and mutation limits an agent needs. The only gap is the reference to user_opinion_source_revision, which is absent from the schema, creating a small traceability mismatch.

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% (only withdraw_user_opinion carries an inline note), so the description must compensate and largely does: it defines agent_verdict as the Agent's analysis, sets user_verdict's default of unreported and the condition for filling it, explains user_note fidelity rules, and clarifies withdraw_user_opinion's required companion fields. The attempt_id and expected_feedback_revision semantics are covered only indirectly via 'current_feedback_revision'.

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 states a specific verb and resource: saving Agent feedback plus follow-up suggestions onto a historical practice attempt, and immediately bounds it ('不能替代自动判分或掌握判定'). It effectively separates itself from siblings by naming read_practice_history and read_practice_feedback as adjacent-but-different tools. It stops short of a 5 only because the tool name ('review_') slightly undersells that the action is a save/write.

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?

Explicit when-to-use and prerequisites: call read_practice_history(include_details=true) to verify the original prompt/answer, then read_practice_feedback to obtain current_feedback_revision (0 on first call). It also gives conditional rules for edge cases — fill user_verdict only on explicit user statements, use withdraw_user_opinion=true only on explicit withdrawal, and read first on version conflict.

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