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

Annotations already fix the safety profile (not read-only, not destructive, closed-world), so the description's job is to add context. It adds the billing behavior ('Debits one credit for authenticated callers'), which the annotations do not carry, and describes the perturbation methodology. It stops short of runtime cost or expected latency, but for a non-destructive analysis call this is solid added value.

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 mechanism, then the when-to-use, then billing, then a concrete example — a sensible ordering with no filler. The embedded JSON example is long relative to the prose but earns its space for a graph-shaped payload.

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?

For a 7-parameter tool with one required argument and a full output schema, the agent has enough: it knows the analysis type, what is swept, what baseline means, the target node default behavior via schema, and the credit cost. Only the boundary against sibling analysis tools is left implicit.

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% and the schema is heavily documented (quantiles, num_quantiles exclusivity, time_period default, target_node_id default all spelled out), so the baseline is 3. The example concretizes the graph shape (root node distribution params, calc node formula, edge wiring, run_params), which helps, but the tool-specific knobs are covered only by the schema, not the prose.

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 and analysis type ('runs a one-at-a-time tornado sensitivity analysis') and then explains the mechanism: sweeping each root node over quantiles with others held at baseline, ranking by impact on the target node. That mechanism-level detail distinguishes it from siblings like scenario_analysis, estimate, and run_simulation without the agent opening any schema.

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 a clear decision rule: 'Use this when you need to know which uncertain input most drives an outcome, so you know what to de-risk or estimate more carefully first.' That is a real when-to-use signal, but it never names the sibling analyses (scenario_analysis, estimate) or states when this tool is the wrong choice, so the agent must infer the boundary itself.

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