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

Annotations declare readOnlyHint=true (safe read operation) and openWorldHint=false. The description adds important behavioral context: the function returns fair-value ODDS BAND, and 'Without offeredOdds it returns fair value + band only (no verdict)'. It also transparently states that results come from the PredictionMarketsPicks desk with a published track record. The one implicit behavioral trait not disclosed is whether the trueWinProbability must be provided externally (vs computed), but the description is otherwise transparent about what inputs and outputs mean.

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?

The description is front-loaded with the core purpose in the first sentence. It then adds parameter semantics, usage guidelines, and provenance in a logical progression. It is slightly longer than necessary (the track-record URL in the last sentence could be a footnote), but every sentence earns its place by adding value. Minor deduction for the slightly dense structure with multiple parenthetical asides.

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?

With 3 parameters (schema 100% coverage, 2 required), no output schema, and a moderately complex tool (fair value grading, band output, optional verdict), the description provides enough context to select and invoke the tool correctly. It covers inputs, output behavior (with/without offeredOdds), the 7-tier verdict categories, and the tool's origin. The only gap is that without an output schema, the description doesn't fully describe the return structure (e.g., the tiers and their thresholds), but for a 'grade versus fair value' tool, the inputs and outcomes are reasonably well described.

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 baseline is 3, but the description adds meaningful context beyond the schema: it explains that legPrices are 'used only for the theoretical assemble ceiling', that offeredOdds can be 'American odds (e.g. -150, 988) or a decimal payout multiplier (e.g. 10.7)', and that trueWinProbability is a 'correlation-aware estimate of the true joint probability'. This adds semantic richness that the schema's generic descriptions lack.

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 uses specific verbs: 'Grade' a 'same-game combo' against 'fair value' on a 'prediction market'. It clearly distinguishes this from siblings like 'calculate_ev' or 'base_rate_gap' by mentioning correlation-aware joint probability and the 7-tier verdict. The resource (combo edge) and action (grading vs fair value) are explicit and sibling-unique.

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?

The description gives explicit when-to-use: 'is this combo worth it', 'grade my parlay quote', 'same-game combo value'. It also provides a explicit when-not-to: 'never grade EV off the product of the leg prices, which no venue pays'. It names an alternative implicitly by referencing 'Kalshi combo RFQ or an SGP product' as context, but more importantly, the sibling list includes 'calculate_ev' which this description avoids competing with by focusing on graded verdicts vs raw EV.

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