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

Annotations already declare readOnlyHint, idempotentHint, openWorldHint=false and destructiveHint=false; the description reinforces this with "answers are not stored or logged, so it is safe to retry," which is the retry-relevant consequence rather than a restatement. More importantly it discloses a major behavioral branch absent from annotations: crisis or violence language short-circuits the normal reading and returns helpline numbers instead. That is exactly the kind of beyond-schema behavior an agent must know before calling.

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 outcome the tool produces, then constraints, then the crisis branch, then the read-only note, which is a sensible priority order for an agent skimming. It is a single dense block rather than scannable segments, and the parenthetical input list (goal, "I" point, obstacles, fears) partly duplicates the schema. Still, nearly every sentence carries decision-relevant information.

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 does not need to re-document return values, and it instead describes only the shape of the payoff plus the exceptional crisis return. Combined with the annotation coverage and the privacy/retry note, an agent has everything needed to call this correctly, handle the fallback branch, and know that retrying is harmless.

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%, so the baseline is 3 and the schema already documents every field, including lengths, array bounds and the language enum. The description still adds real semantic constraints on what may be passed: only the user's own short phrases, and no diagnoses, symptoms, medication or health details. That is meaning beyond the schema's type and length rules, though it does not clarify field-level distinctions the schema leaves implicit.

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 and artifact: builds a Ladder from the user's own answers and returns zone-grouped steps, an interpretation, two reflection questions and a link to the full diagnostic. The enumeration of inputs (goal, "I" point, 3-5 obstacles, up to 5 fears) matches the schema exactly, so the agent knows precisely what it produces and from what. Placing it "after the user has answered the self-check questions" functionally separates it from get_ladder_questions.

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 an explicit trigger (once the user has answered the self-check questions), explicit exclusions (do not pass diagnoses, symptoms, medication or other health details, because the tool does not analyze them), and a distinct fallback path for crisis or violence language. It also tells the agent what to do when that fallback fires: show only the crisis message and helpline numbers, with no steps and no reading. The sibling get_ladder_questions is not named outright, but the sequencing plus the negative guidance leaves little ambiguity.

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