suggest criteria

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

Extremely rich for a tool whose annotations only state readOnly=false, openWorld=true, idempotent=false, destructive=false. The description discloses billing (SPENDS MONEY, billed under assist:suggest), auth requirements (OWNER/ADMIN, evals:write scope), error conditions (400s with counts), the valid body-less POST, dedup behavior (same-name skipped, never overwritten), the trust model (drafts never gated until human review), and maxDuration 300s. No contradiction with annotations — readOnly=false and idempotent=false align with a billable draft-creating call.

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?

Information-dense and front-loaded: the core purpose leads, followed by endpoint/scope, cost warning, return shape, and edge cases. Every sentence earns its place. The only minor redundancy is the billing warning appearing twice ('SPENDS MONEY' and 'SPENDS THE WALLET'), which can be read as deliberate emphasis on a critical fact.

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?

No output schema exists, so the description carries the full burden of explaining return values — and does so exhaustively, including the exact criterion object shape plus critiques_used and skipped_duplicates. Prerequisites, failure modes, auth, timeout, and the dedup/trust behavior are all covered. Nothing essential is missing for an agent to invoke this correctly.

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 100% — judge_model is fully documented with its default model, the playground-catalog requirement, and whitespace-only fallback. The description adds the body-less POST default-judge behavior, but since the schema already carries the parameter meaning, the baseline 3 is appropriate.

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?

States a specific verb (clusters), resource (failure critiques → up to 5 DRAFT judge criteria, one per failure mode), and the workflow position (after grading fails with reasons). The draft/assist-suggested nature clearly distinguishes it from siblings like create_criterion, update_criterion, and auto_improve_criterion without needing to open their schemas.

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?

Gives clear context: use it after grading a batch of fails with reasons to bootstrap criteria that are then reviewed and aligned. It also states the prerequisite (≥10 FAIL grades with critiques, only the 200 most recent considered) and follow-up workflow. However, it never explicitly names alternatives such as create_criterion for hand-writing a single criterion or run_criterion_alignment for the post-draft step.

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