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

With no annotations provided, the description carries the full burden — and it delivers richly: missing horizons are "skipped, never zero-filled," funding/OI/basis "are not a served tape here and are not invented," stale data "is said on the row, name stays, no number invented," return size is capped ("RETURNS ≤20"), ranked_count semantics are defined, and safety state flags are disclosed (do_not_place_from_mover_rank, edge=no_measurable_edge, not_a_forecast=true). It even states the auth mechanism (X-API-Key). This goes far beyond what annotations would typically provide.

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 dense and every clause adds value (universe definition, signal list, data-handling rules, output fields, safety flags), so there is little fluff. However, it is a single ~300-word run-on paragraph with heavy ALL-CAPS emphasis and stacked parentheticals; breaking it into scannable sections (input / output / safety) would materially improve parseability. The content earns the length, but the structure does not.

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?

Given the tool's complexity (scan + rank + sleeve with many behavioral caveats) and the absence of both output schema and annotations, the description is exceptionally complete. It defines the input universe and its endpoint, the exact signal bundles and their exclusions, missing-data conventions, every key output field (blendedScore, decile, decile_convention, as_of, signal-bundle keys, ranked_count), the ≤20 cap, the ordering-only (non-forecast) semantics, and the auth requirement. An agent could invoke and interpret this tool correctly with no further information.

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 description coverage is 0%, so the description must compensate — and it does. "SLEEVE a bounded top-N (default 20)" explains sleeve_n; "$1M 24h floor" maps to min_dollar_volume (default 1,000,000); "crypto desk" aligns with asset_class's only enum value. The mapping is inferable but not explicit — parameter names are never mentioned, so an agent must bridge the prose to the schema fields, which keeps this just short of a 5.

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 a specific verb-resource pair: "ONE sitting contract for the crypto desk: SCAN the full liquid USD book ... RANK every liquid name ... SLEEVE a bounded top-N." It clearly distinguishes itself from siblings — the full book stays on /api/crypto/universe, it is the crypto counterpart to tengu_equity_sitting, and it contrasts its universe definition against "Alpaca's 36" and "a pre-cut 20."

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 strong context on when to use this tool: use it for a ranked, sleeved view of the liquid crypto book, and go to /api/crypto/universe if the full book is needed ("Sitting GET still RETURNS ≤20"). It also excludes alternatives via safety flags ("not_a_forecast=true", "FIRM never places"). However, it never explicitly names sibling tools like tengu_crypto_overnight or tengu_ml_top_picks as alternatives, leaving some comparative guidance implicit.

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