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

Goes well beyond the annotations (readOnlyHint=false, idempotentHint=false, destructiveHint=false) by disclosing what is actually written and when: include_claim_link 'stores a small record for 30 days (only when enabled)'. It also exposes domain behavior an agent could not infer: FDA-label checking with 'major picks dropped, moderate flagged,' the warning that 'unlisted medications are NOT checked,' avoid-level cautions leading as cautions_summary, and the serious-disease no_match path. No contradiction with the annotation profile.

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-loading is good — stack size and scope come first — and nearly every clause carries actionable information. However the whole definition is a single semicolon-and-parenthesis-choked run-on, making it hard to scan; the interaction, caution, and claim-link rules would parse far better as separate structured sentences.

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?

There is no output schema, yet the description enumerates the return shape precisely: 'Each item: why, grade, study dose (context), PMIDs, cited cautions... same-mechanism and interaction flags,' plus data.important_cautions and data.claim.claim_url. For a composite, non-idempotent tool this covers purpose, inputs, outputs, failure modes, and follow-up routing.

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 the baseline is 3, but the description genuinely extends the schema: `exclude` 'also takes medications by generic name (checked against FDA labels)' and unlisted meds are explicitly NOT checked, and `include_claim_link` is explained as storing a 30-day record and returning data.claim.claim_url. It adds interaction semantics and side-effect meaning the schema does not carry.

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?

States a specific verb+resource+scope: builds a 'starter stack (default 3, max 5) of recommendable supplements for a goal,' with concrete examples ('insomnia', 'LDL cholesterol'). It distinguishes itself from the list-style siblings by framing output as curated, interaction-checked picks ('never Rx or research', 'hypotheses' for biomarker-proxy rows), though it never names get_supplements_for_condition/get_supplements_for_biomarker as the alternatives it supersedes.

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?

Provides real routing guidance: excludes serious disease ('not offered for cancer or other serious disease (no_match)') and tells the agent to 'ask, then use get_supplement_interactions with `drug`' when a medication is unlisted. Clear context and a named fallback, but it does not contrast against the other recommendation-style siblings (get_supplements_for_condition / get_supplements_for_biomarker), so the when-to-choose-this branch is only partial.

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