PGxAI Andromeda

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

The annotations already establish that this is a read-only, idempotent, non-destructive, closed-world lookup. The description adds useful behavioral context beyond annotations, notably that a population label with no sub-codes returns an empty list and that empty results are expected rather than errors.

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 long category list is necessary, but the definition is a run-on paragraph that mixes this tool's parameter guidance with downstream conditions[] formatting rules. It is front-loaded with the mapping instruction, but the structure could be much clearer and less repetitive.

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?

An output schema exists, so the description need not explain return values in detail, and it does cover key edge cases such as empty sub-code lists and rejected input formats. It is nearly complete, though the overlap between this tool's category parameter and review_medication_evidence's conditions[] requirements could be separated more cleanly.

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 the schema already says category must be an exact top-level name, but the description adds substantial meaning by providing the full category list and examples such as Cardiac disorders and Diabetes. It also clarifies when to pass a category name as-is versus using a more specific sub-code name in the downstream workflow.

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 specific mapping action and explicitly names the sibling tool review_medication_evidence, so the agent can distinguish the workflow. However, it is not a clean statement of what get_condition_category itself does because it focuses on mapping user terms and submitting sub-code names rather than directly saying it retrieves category/sub-code data.

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 gives explicit sequencing guidance: map terms independently before calling review_medication_evidence, prefer the most specific matching name, submit a category only when the topic is general or no sub-code fits, and never submit the icd10_code. It also states exactly which condition string formats downstream conditions[] accepts and which are rejected.

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

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

Annotations already cover read-only/idempotent safety, and the description adds genuinely new behavior: the status vocabulary (findings, checked_no_findings, nothing_to_check, no_data_returned) and the crucial caveat that checked_no_findings is not proof of absence or safety. It also warns that source_url, phenotype and implications may be null, and that severity is nominal rather than ranked.

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?

Purpose and usage lead, with status semantics and caveats after — good front-loading. Some sentences restate what the output schema already carries (e.g. which fields may be null), costing a little tightness in an otherwise dense, well-paragraphed block.

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 a rich output schema, the description supplies the parts the schema can't: status interpretation, inconclusiveness of no_data_returned, and the evidence-preservation instruction. Nothing an agent needs to call or interpret this correctly is missing.

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; the description still adds value by tying conditions to an external lookup tool and by clarifying that drugs may be English names or RxNorm identifiers. It does not, however, explain the confirm_no_direct_identifiers resend flow beyond what the schema already states.

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 first sentence names the specific resources (DDI, PGx, DCI) and the operation (lookup published evidence), so an agent knows exactly what comes back. It also implicitly scopes the tool to multi-item queries (two or more drugs, drug + disease, drug + gene), which separates it from a general drug-info sibling.

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 enumerates the three qualifying query shapes and states the prerequisite workflow — map condition terms through get_condition_category first, using the most specific sub-code and a category only for general topics. It also declares the exclusion that omitting genes means no PGx analysis, which prevents a common misuse.

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