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

Annotations are present but sparse (all false), so the description carries substantial extra value: it discloses the fee model, deterministic seeding, the return payload (script/plan/data fingerprints), a query quota (8 per day per client, separate public pool), and explicit failure behavior when the engine is unavailable. This meaningfully exceeds what the annotations convey, though it does not explain runtime or latency expectations.

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?

It is a single dense paragraph that front-loads the core operations before parenthetical detail, and every sentence carries information. The heavy nesting of parentheticals and mixed concerns (engine, fees, quota, output, failure) make it slightly harder to scan, but there is little pure filler.

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 2-parameter tool with an output schema, the description covers the engine, determinism, fees, quota, and failure modes, so an agent has most of what it needs to invoke it correctly. The absence of explicit contrast with the closest sibling tools (backtest_arenascript, monte_carlo) is the main remaining gap.

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 only 50% (holdout_percent has no schema description), and the description compensates by defining its bounds and the special case of 0 (no holdout, Monte Carlo degrades to monte_carlo). The only schema-ignored param is holdout_percent itself, so the description adds real meaning beyond the structured fields.

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 names the specific composite operation (backtest + chronological holdout + Monte Carlo) and the exact engine that performs it, with concrete execution details (same Rust binary, 10 bps per side, deterministic seed from trade fingerprint). It partially distinguishes itself from siblings by referencing monte_carlo, but does not explicitly contrast with backtest_arenascript or backtest_strategy, leaving the agent to infer the difference.

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?

Usage is only implied: the holdout_percent semantics (10-40, or 0 meaning Monte Carlo becomes equivalent to monte_carlo) hint at when the tool is appropriate, but there is no explicit 'use this instead of X when...' routing against the many related backtest/monte_carlo siblings. The closing 'Descriptif, pas un KEEP' is context-specific and not a genuine usage guideline.

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