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

The annotations already establish read-only, idempotent, non-destructive behavior, and the description goes well beyond them with ordering guarantees, duplicate collapsing, notInDataset and omittedForSize buckets, and rejection of malformed or bad-check-digit UIDs. It also explains that absent vatRegistration means enrichment has not run and that ANALYSED does not guarantee every field, which is exactly the kind of semantic caveat annotations cannot express. Slightly dense but creditable.

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 with the core purpose, then the caveats in rough priority order. Several sentences are em-dash-heavy and pack three clauses each, but each still carries distinct information, so nothing is padding.

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 one-parameter read tool with an output schema and full annotations, this covers nearly everything an agent needs: batching limits, partial-result signaling, ordering, and failure semantics. The only gap is not spelling out where the UIDs come from, which the sibling set handles implicitly.

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% and there is a single required parameter whose format (formatted or plain, max 25) is already documented in the schema. The description adds behavior around the parameter — request order preserved, duplicate UIDs collapsed to one, malformed values rejected — but nothing about UID syntax or what triggers a bad check digit. Baseline 3 is appropriate.

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?

"Full detail for one or several companies (1 to 25) by UID" names a specific verb (retrieve full records), the resource (companies), the key (UID) and the batch bound. The phrase "the second half of a search, without a call per row" positions it precisely against search_companies and load_company, so an agent can pick it without opening a schema.

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 context is clear: it is the follow-up to a search and the batch substitute for per-row calls. It also names load_company as the fallback when a UID is missing from the dataset. What is missing is an explicit pointer to the sibling that produces the UIDs (search_companies / find_company_by_name) or a stated "do not use for name lookup" exclusion.

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