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

Annotations cover the safety profile (readOnly, idempotent, non-destructive, closed-world), so the bar is lower. The description still adds real behavioral context beyond them: transient processing with no storage/logging/retention, the auto/server/browser execution modes with gpu:true always delegating, and an exported AP2 artifact carrying execution_hash for provenance. What is missing is any statement of what the computed result looks like despite the artifact export.

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?

Front-loading is fine and the operational content is relevant, but there is noticeable redundancy ('compute node (compliance_mandate). Deterministic OpenChainGraph compute node') and the sentence about the FV-status snapshot/offline receipt is a tangent that does not help an agent call the tool. Roughly a third of the text is chain boilerplate.

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?

An output schema exists, so return-value description is not required, and the chain wiring, execution modes, and data-handling policy are covered well. The gap is domain completeness: with 4 optional params, no required fields, and 80+ sibling calculators including compute_rbc_action_level_private, the description never says what this tool computes relative to those, leaving selection ambiguous.

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 100% and the description largely restates the schema's own compute-mode wording ('auto' = server for gpu:false with registered kernel, 'browser' = delegation URL). It adds nothing on parent_hashes/parent_tool_ids beyond what the schema already says, and policy_parameters field names are explicitly deferred to an external manifest. 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?

The title/name identify it as a NAIC RBC Action Level calculator, which is a specific resource, but the description body never explains what an action level is or what the calculation decides; it is dominated by OpenChainGraph compute-node mechanics. It also restates the title verbatim (name + 'compute node'), which is tautological. An agent can tell it is a calculator but not precisely what it returns.

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?

It gives an explicit input constraint ('Use synthetic or anonymised inputs only') and places the tool in a chain via upstream/downstream artifact ids. However, it gives no when-to-use vs. when-not guidance, and notably does not distinguish it from the close sibling compute_rbc_action_level_private or the other RBC/exposure calculators. Usage is only implied by the chain wiring.

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