Stipple — Reference Verification

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

The readOnlyHint annotation is present, and the description adds substantial behavior beyond it: server-side fetching, the three default checks (provenance, citation resolution against arXiv/Crossref, internal math recomputation), the deep=true mode with a time estimate (~a minute), the return shape (trust summary, per-item tables, permalink), and an explicit 'HONEST BOUNDARY' disclaiming that it reports coverage, not truth. This is model-level transparency.

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?

Despite its length, every sentence earns its place: purpose, usage triggers, input modes with tradeoffs, default-vs-deep behavior, returns, and limitations are each in their own paragraph with helpful flags like 'USE THIS WHEN' and 'HONEST BOUNDARY'. The description is front-loaded with purpose and contains no filler or repetition of the schema.

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 the complexity (five parameters, three mutually exclusive input modes, no required params, zero schema coverage), the description is remarkably complete: it explains the mutual-exclusivity constraint, the deep-mode toggle, time cost, outputs, and interpretatin caveats. The output schema exists, so return values need not be spelled out. The only substantive gap is the undocumented filename parameter and the absence of any error-mode notes (e.g., failed URL fetch or malformed PDF).

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 0%, so the description carries the full burden, and it compensates well: url (public HTTP(S), text (pasted markdown/plain prose), bytes_b64 (base64 PDF with exact link-annotation URLs), and deep (default fast vs fetch-each-source mode) are all given meaningful semantics, including cost tradeoffs. However, the filename parameter is never mentioned anywhere in the description or schema, leaving one of five parameters undocumented.

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 opening line, 'Fact-check a document's REFERENCES and CLAIMS — built for AI-generated reports whose citations must be checked before they're trusted,' states a specific verb (fact-check), a specific resource (references and claims), and the target use case. There are no sibling tools to differentiate from, and the title 'Verify a report's citations' is reinforced rather than merely restated.

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 includes an explicit 'USE THIS WHEN' block with concrete trigger scenarios and example user questions ('is this accurate / legit?', 'did the AI make this up?'), plus proactive guidance ('use it proactively before relying on any AI-written document'). It also routes input selection by explaining when to choose url vs text vs bytes_b64, so an agent knows how to invoke it correctly.

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