Reference Checker: Fake, AI-Hallucinated & Retracted Citations

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

Annotations already declare readOnlyHint=true, openWorldHint=true, destructiveHint=false and idempotentHint=false, and the description is consistent with all of them ('read-only'). It adds real value beyond them: the Apify actor and account it runs under, $0.01-per-reference billing, the run_max_charge_usd cap, that finding nothing is not charged, timeout behavior (5-15 s typical, raise run_timeout_secs to 300), and that monitor mode persists state in a key-value store. That is substantive operational context, though it is packed into a single dense block rather than organized by concern.

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 sibling differentiation are front-loaded, followed by usage, then parameters, then runtime/billing behavior, then return shape — a sensible ordering. The prose is dense but nearly every sentence carries non-obvious information (billing mechanics, monitor state, timeout escalation) rather than restating 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?

For a 14-parameter, zero-required, billed, open-world tool with no output schema, the description covers what an agent needs: at least one of bibliography/dois/titles is required, cost and cap mechanics, timeout escalation, persisted monitor state, and an explicit per-item return field list (status, verdict, matchedDoi, matchScore, mismatches, integrity.status, openAccessUrl, warnings). Little is left to inference.

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 schema already documents all 14 parameters and the baseline of 3 applies. The description does add cross-parameter meaning the schema cannot: lowering minMatchScore trades recall for false matches, includeCandidates is free, run_max_items fills maxReferences when unset, and setting maxReferences higher can bill more than is returned — genuine semantic value, but scattered and not systematic.

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 sentence gives a specific verb and resource ('Verify an existing bibliography against Crossref') plus the failure classes it detects (fake/hallucinated, miscited, retracted) and the output shape ('one row per reference'). It then explicitly contrasts itself with citation_formatter ('judges whether a reference is real and accurate; it does not produce formatted styles'), so an agent can separate it from siblings without reading schemas.

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 states the trigger ('Use when you must audit a reference list, DOIs or titles before submission or review') and both exclusions with named alternatives: discovery goes to literature_search or pubmed_evidence_table, restyling goes to citation_formatter. Monitor-mode usage (reuse the same monitorName on scheduled runs) is also specified.

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