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

With no annotations to fall back on, the description does the full disclosure. It reveals that the tool forces: that leg cost depends on a 'turns' multiplier, that at turns=1 the full round trip is charged while at turns=50 it approaches zero, and that the old 'venue recommendation' output was measured and deliberately removed. It does not explicitly mention that the tool makes live API calls or its latency, but the explanatory depth and warnings about not overloading turn=1 as a per-trade cost show it goes beyond the bare definition.

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?

While longer than average, it adapts to the tool's conceptual complexity. It introduces the 'point' with a vivid phrase (the cost of the currency you were never quoted in), but the structure is logical: purpose, examples, caveats, parameter definitions, and a closing API-key note. It is a single cohesive block; the front-loading of the one-sentence 'POINT' is strong. A few supporting sentences could be trimmed, but not enough to break the value.

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 tool with an output schema and several behavioral caveats, the description covers all key context: the motivation (why the leg is invisible in other TCA), the inputs, the output's intended meaning (leg_as_pct_of_execution_cost), and explicit counterfactual warnings (it does that, not that). It also notes the Pro API key requirement, which is not in the schema. The only missing piece is the exact response data model, but the output schema covers that.

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?

Input schema is minimal – only names, no descriptions (coverage 0%). The description fully compensates: 'coin' is a ticker ('sol'), 'notional_usd' is order size in USD, 'side' is 'buy' or 'sell', and 'turns' is explained as how many times the desk repatriates to USD (with concrete thresholds). It even clarifies the default of 1 turn and warns against quoting the turns=1 figure as a per-trade cost for a continuously trading desk.

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 states a clear, specific purpose: to compute the hidden currency-conversion leg cost for a trade, using a finite live order book, and to re-rank venues including that leg. It also explicitly disclaims venue-recommendation, which distinguishes it from sibling tools like route_order and leg_rates. Terms like 'walks the live USD/USDT and USD/USDC books' and 're-ranks venues on cost INCLUDING the leg' make the action and resource unmistakable.

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?

It provides concrete example queries (e.g., 'what does it really cost me to trade $5M of SOL') and an explicit exclusion: 'IT DOES NOT TELL YOU WHICH VENUE TO TRADE ON, and you must not present it as if it does.' This is stronger than typical guidance because it states both when to use and when to avoid misusing the tool, and references the output field leg_does_not_change_venue that encodes the non-decision.

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