Papers

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

Annotations already cover readOnly, idempotent, non-destructive and openWorld, so the safety profile is settled. The description adds genuinely non-obvious behavioral context: rows correspond one-to-one with API hits, and an empty result reflects an empty API response rather than proven absence of citations. It does not mention the limit cap or rate-limit behavior.

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?

Three short sentences, zero filler, with the primary purpose and the two relation modes front-loaded before the caveat.

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 return shape needn't be restated, and the description goes further by explaining empty-result semantics. The main remaining gap is the absence of guidance on the limit parameter and on pagination beyond the 20-row cap.

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 must carry the param burden. It explains the critical relation enum well (cited_by = papers citing the work, references = papers the work cites), but identifier and limit are left entirely to the schema, which supplies only type and bounds, no meaning.

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 gives a specific verb (List) and resource (papers) plus the exact selection criterion, and it defines both relation modes (cited_by vs references) inline. It is immediately clear this is a citation-graph lookup rather than a keyword search, though no sibling tool is named explicitly.

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 tells the agent which relation value maps to which direction of citation, which is actionable parameter-selection guidance. However, it never says when to prefer this over search_papers or get_paper, and it gives no prerequisites or exclusions.

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 declare readOnly=true, idempotent=true, destructive=false, so the safety profile is covered. The description goes beyond them by disclosing normalization behavior (titles kept as the source provided them, display names split mechanically so author particles may need human review) and what to do on resolution failure — genuinely useful traits not derivable from structured fields.

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?

Three tight sentences with the core action front-loaded and no filler. Slightly dense but every clause carries 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 return-format explanation is rightly omitted, and the tool is read-only so no destructive caveats are needed. However, for a tool whose only real input ambiguity is the identifier, leaving its accepted forms unstated leaves a meaningful gap.

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. It enumerates the style values (matching the enum) but never says what form "identifier" takes — DOI, arXiv ID, URL? — which is the single most important thing an agent needs to construct a correct call.

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 (format) plus resource (a real paper) and enumerates the four supported output styles, which cleanly separates it from the sibling read/lookup tools (find_related_papers, get_paper, search_papers). An agent can identify the tool's job without inspecting the 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?

It provides a failure-mode rule ("If the paper cannot be resolved, do not hand-write a citation"), which is useful conditional guidance, but never states when to prefer this tool over its siblings or that an identifier must first be obtained. Usage is only implied.

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 declare readOnly, idempotent, non-destructive and openWorld, so the safety profile is covered. The description adds genuinely new behavioral context: metadata provenance ('the API that resolved the identifier') and the failure contract ('if no source returns a record, say so. Do not guess the abstract'), which is exactly the kind of anti-hallucination instruction an agent needs.

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?

Four short sentences, front-loaded with the operation and its inputs; each sentence carries distinct information (what it returns, identifier types, provenance, failure handling). Slightly padded by listing the fields, which the output schema already provides.

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 return-value explanation is not strictly required, and annotations cover safety. The description fills the remaining critical gap — what to do when resolution fails — making it essentially complete for a single-identifier lookup tool.

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?

There is one parameter with 100% schema description coverage, and the description only restates the identifier vocabulary already documented in the schema. It adds no syntax, precedence, or format guidance (e.g. how to disambiguate an ambiguous bare string), so the baseline of 3 applies.

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?

Specific verb ('Fetch') plus a single, clearly bounded resource ('one paper'), and it enumerates the returned fields (title, authors, year, venue, abstract) and the accepted identifier types. It is unmistakably narrower than search_papers or find_related_papers, but it never names a sibling to make the contrast explicit.

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?

Usage is implied by the accepted identifier types — you call it when you already have a DOI/PMID/arXiv/OpenAlex id — but the description never states when to prefer this over search_papers or find_related_papers, nor any prerequisite or exclusion.

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 declare readOnly, idempotent, openWorld and non-destructive, so the safety profile is covered. The description adds real behavioral context beyond them: results are limited to what upstream APIs returned, each carries a source link and identifier, an empty list is a legitimate signal, and the model must not fabricate entries.

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?

Four tight sentences, front-loaded with purpose, then filters, then return semantics, then the anti-fabrication warning. No sentence is redundant and nothing is buried.

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 return shapes need no explanation, yet the description still clarifies the source-link/identifier payload and the empty-result meaning. Combined with rich annotations, an agent has everything needed to call this correctly.

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?

With schema coverage at 57%, the description compensates well by naming the filters (year_from, year_to, field, open_access, source), enumerating the source values, and explaining that source=all mixes indexes. Only 'limit' and the semantics of year bounds are left solely to the schema.

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 resource ('Search peer-reviewed and preprint records') plus the input mode (question or keywords), which clearly distinguishes it from siblings like get_paper and find_related_papers, which operate on a known paper. An agent can pick this for open-ended discovery without opening the 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?

Usage is implied by the filters and the source list, and the closing line ('Do not invent papers to fill a gap') gives a genuine anti-hallucination directive for empty results. However, it never states when to prefer this over find_related_papers or get_paper, so sibling routing is left to inference.

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