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

Goes well beyond the readOnly/idempotent/non-destructive annotations: discloses ordering (most recently filed first, nothing else), the snapshot cutoff date, that subsection input is stripped and reported in subsection_dropped, that section_in_snapshot=false marks an absent section shown as written, and that every response including 400/503 carries note and coverage_note. It also quantifies reliability (precision 0.9845 on a blind sample, recall unmeasured) and names extraction gaps, which is exactly the context an agent needs to phrase confidence.

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 and purposeful in its first two sentences, but the body is a dense statistics block (25,093 of 60,444 sections, 10,745,929 records, occurrence-level 0.8870, distinct-pair 0.1697, Wilson intervals) that partly duplicates what the description itself says the sources tool provides. Several figures earn their place (base rate for empty results), but the precision/recall statistics could be trimmed without loss for a typical calling agent.

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 no output schema, the description carries the burden of describing returns and does so for the non-obvious parts (parentheticals[] verbatim with 95.8%-empty base rate, section_in_snapshot/snapshot_detail, always-present note and coverage_note). It does not describe the shape of the citing-opinion records themselves (fields, whether cluster/opinion identifiers are included), which is the remaining gap for a caller building on results.

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 100% and each parameter is documented in-schema, so the baseline is 3. The description adds little parameter-specific meaning beyond what the schema already says — subsection stripping, the title+section alternative, and the limit/offset caps all appear in the schema text — though it does explain what happens at the semantic edges (dropped subsection reporting, refused state citations).

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?

Opens with a specific verb and resource ("Lists the opinions whose text cites one U.S. Code section") and immediately scopes it ("Federal only"), which separates it from sibling statute/statute_search tools that operate on statutes rather than citing opinions. It also states the semantic limit of an entry (text names the section, says nothing about treatment), so an agent knows what a hit does and does not mean.

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 explicit when-not guidance (state citations are refused with a visible error) and warns that zero results is the majority outcome, so an empty list is not evidence of no citations — a rare and decision-relevant usage caveat. It also routes to the sources tool for the full coverage statement. It stops short of naming when to prefer statute_search or statute for resolving the citation string first, so it is clear context rather than a complete routing rule.

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