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

Rich beyond the annotations: it discloses three concrete deviations from the model's firing rule (no place-capacity enforcement, only the heaviest parallel arc kept, inhibitor read as a read arc) and a per-call cost of 2 tokens. The only friction is the 'the read sim_calibrate' framing against readOnlyHint=false / destructiveHint=true; this is plausibly attributable to token spend rather than model mutation, so it reads as a mild tension rather than a contradiction, but it is not fully reconciled in the text.

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 the purpose and the sim_calibrate contrast, then moves to requirements, replay semantics, caveats, and cost — a sensible order. It runs as one dense block with some run-on phrasing ('offered on its own and in full'), but nearly every clause carries operational information.

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?

There is no output schema, so the description must carry the return contract and it does: the four metric families, per-trace diagnostics naming activities that could not fire, and the residual block with shortfall position, count, and a where-fidelity-holds summary. Requirements, caveats, and cost are all present, so an agent can call it correctly without further discovery.

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%, so the bar is baseline 3, and the description then earns extra: it specifies the log format (CSV with case_id, activity, timestamp), that activities must be transition ids, and that this is the shape sim_dataset emits. That meaningfully constrains how the `log` argument must be built, which the one-line schema description does not.

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 precise verb and resource ('check how well a stored model matches an observed event log') and explicitly distinguishes itself from sim_calibrate by being the read-only scoring half rather than the rate-rewriting half. It also enumerates the concrete outputs (fitness, precision, generalization, simplicity, residual block), so the agent knows exactly what it is being asked to run.

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?

Names the alternative (sim_calibrate) and the condition that selects this one: scoring without rewriting rates. It goes further with when-not guidance — the model must be the per-case workflow, because a resource net with places shared across cases will not fit — and specifies timestamp-order replay semantics.

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