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

Annotations only carry readOnlyHint=true, idempotentHint=true, destructiveHint=false. The description adds substantial context beyond these: 'Inputs are processed transiently... not stored, logged, or retained', deterministic execution, browser-delegation behavior for gpu:true nodes, and the provenance export ('AP2 artifact with execution_hash'). No contradiction with annotations.

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-loaded with core facts (verb, resource, deadline, determinism) and logically organized, but verbose. The browser URL and the extended FV-status receipt semantics ('a snapshot, not a subscription... verifies offline regardless of whether that file is ever fetched') are operational metadata that add noise for tool-selection purposes.

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?

Thorough for a complex compute-node with no output schema: covers compute routing, transient data handling, provenance chaining, regulatory deadline, downstream feeds, and verifiability. The main gap is that policy_parameters field names are only deferred to a manifest, and exact return shape beyond the AP2 artifact is unspecified.

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%, so the baseline is 3. The description reinforces the compute parameter behavior (detailing auto/server/browser routing) and references the AP2 artifact that maps to parent_hashes, but it adds no per-parameter semantics beyond what the schema already documents; policy_parameters is deferred to a manifest in both schema and description.

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 action ('replays supervisory scenarios') on a specific resource ('OpenChainGraph compute node (capital_assessment)'), with a clear regulatory identity ('DFAST-lite', 'Annual re-pin — Fed publishes new supervisory scenarios each February'). It is clearly distinct from siblings like run_stress_test_scenarios or compute_stress_test_scenarios by being a deterministic replay with chain provenance and compute-mode routing.

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?

Explains when each compute mode applies ('auto' server-side on Cloudflare Workers for gpu:false with registered kernel; 'browser' returns delegation URL; gpu:true always delegates) and imposes a hard input constraint ('Use synthetic or anonymised inputs only'). It does not explicitly name a sibling alternative for when NOT to use this tool, but the mode-routing guidance gives an agent clear call-time direction.

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