LeakData

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

Goes well beyond the read-only annotation by explaining interpretive limits: a matching prefix is only a group-level aggregate, not confirmation of exposure; an empty group does not prove a password is safe; and the tool returns aggregate values only. These are critical semantic warnings the agent needs to avoid overstating results.

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 description is front-loaded with the condition for use and every sentence adds safety or scoping guidance. There is no filler; warnings about inventing prefixes, accepting plaintext, and misinterpreting results are each distinct and necessary. The length is justified by the safety-critical nature of the tool.

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 one required parameter, a rich schema description, output schema present, and read-only annotations, the description covers the remaining gaps: when to avoid calling the tool, how to handle general password questions, and how to phrase results responsibly. No important usage context 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 adds meaningful provenance constraints beyond the schema: the prefix must be locally computed by the user, must not be invented or reused, and plaintext passwords must never be accepted. These are behavioral constraints rather than format constraints, adding real value.

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 title and opening directive ('Call this only when... a five-character SHA-1 hash prefix') make the tool's trigger condition clearociative. The description repeatedly frames the tool as reporting aggregate values for a prefix, which conveys the core purpose. However, it never states in one explicit sentence that this tool looks up a hash prefix against breach data, so the primary action is more implied than directly named.

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?

The description gives explicit when-to-call conditions: only for a locally computed five-character SHA-1 hash prefix from the current message. It also gives clear exclusions: never for plaintext passwords, never for earlier or invented prefixes, and never for general password questions, where local computation guidance should be provided instead.

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 only give readOnlyHint=true, openWorldHint=false, destructiveHint=false. The description adds: all calls are live (never report earlier results as a new check), failures must be reported as check-incomplete, the server resolves the verified primary email, and the tool does not return arbitrary identifiers, usernames, phones, raw leak records, or credentials. These are important trait disclosures far beyond the annotations.

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 description is long (~150 words) but front-loads the primary rule and most decisions. Every sentence is a real operative rule, yet there is some redundancy: 'Never present earlier conversation results as a new check' is partially re-stated by 'Never guess an email address or recover it from earlier conversations'. That normalization is justifiable for a security-sensitive tool, so a 4 rather than a 5.

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, but the description otherwise covers all needed call-time knowledge: the type enum behavior, query omission/preservation for ownership validation, server-side resolution, failure reporting, what is not returned, and the only one available non-call case (domain checks). For a security-critical tool with one enumerated type and one optional query, nothing needed to invoke it 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 50% (only the `type` parameter's enum and `query`'s basic description in the schema). The description materially adds: set `type` to 'email', set `query` only if this tool has been tested with a user-supplied email for ownership validation, omit it otherwise, and never infer an email from earlier conversations. This fully compensates for the schema's gaps for the named behavior.

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 is not a tautology: it states a specific action ('check their verified primary email'), a specific resource (the linked LeakData user's verified primary email), and the trigger condition ('when the linked user asks'). It also implicitly distinguishes itself from the sibling leakdata.password_prefix_check by framing its scope as email verification, which an agent can differentiate without opening the sibling's 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?

Gives explicit when-to-call, when-not-to-call, and exclusion rules: 'Use this when the linked LeakData user asks to check their verified primary email', 'Domain checks are not available through this app; explain this without calling a tool', and 'if the call fails, say the check could not be completed'. It even defines the exact parameter strategy ('set type to email and omit query' vs. 'preserve it in query'), leaving no decision to inference.

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