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

The description extensively discloses behavioral traits beyond the annotations: it is deterministic, explains compute modes (auto/server/browser), states inputs are processed transiently and not stored/logged/retained, mentions the AP2 artifact export with execution_hash for chain provenance, and provides a URL and FV-status link. No contradiction with annotations (readOnlyHint, idempotentHint, etc.) is present. This is a model of transparency.

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 description is ~250 words and dense with technical details (compute modes, data retention, artifact, FV-status). It front-loads the purpose and includes operational details, but it is verbose and could be tightened by moving less-critical boilerplate (e.g., FV-status explanation) to a secondary section. Structure is adequate but not optimally concise; every sentence carries information, but efficiency is moderate.

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?

Given the tool's complexity (compute modes, artifact export, provenance) and the absence of an output schema, the description covers most runtime behavior: compute execution contexts, data retention, artifact generation, and a verification link. However, it does not describe the actual return payload (e.g., whether the tool returns a boolean, a risk assessment, or just the artifact). This gap is significant for an agent that needs to interpret the result, though the artifact export likely implies the response contains the execution_hash and possibly computed values. Overall, it is nearly complete but lacks explicit output semantics.

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?

The input schema has 100% coverage—every parameter (compute, parent_hashes, parent_tool_ids, policy_parameters) has a description. The description's text adds little beyond what the schema already states; for example, it references the manifest for policy_parameters field names. The description does not introduce new meaning or clarify parameter usage beyond the schema, so it meets the baseline for high schema coverage.

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 clearly states the tool's purpose: a compliance check for securitization risk retention, within an OpenChainGraph compute node context. However, it does not explicitly differentiate this from the many sibling 'check_*' tools (e.g., check_conforming_loan_limit, check_credit_concentration_topn_sector), leaving the agent to infer uniqueness from the specific resource name. The verb 'check' and resource are present, but sibling differentiation is absent.

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 description provides some usage context: it is a compliance_mandate compute node, suggests using synthetic or anonymised inputs, and mentions compute modes. It also notes that gpu:true nodes always delegate to the browser. However, it does not specify when to choose this tool over alternative checks or provide exclusions (e.g., 'if you need X, use Y instead'). There is no guidance on selecting among sibling tools, so the agent must rely on the resource name to infer applicability.

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