agentleFS
surendranbsurendranb/find-research-papers-mcp/skills/interpreting-errors/SKILL.md

interpreting-errors

Skill7 starsChanged 54 days agoMITView on GitHub

How to read this server's error and skipped shapes (search_papers, get_paper) and recover

  • In this repository2 other files it gives its agents.

The file

Files
SKILL.md

Interpreting errors from find-research-papers-mcp

Factual guide to every error shape this server produces, and the recovery move for each. Errors here are structured data, not crashes — read them.

search_papers

A search never hard-fails on a single source. Failures appear in skipped:

  • {"source": ..., "reason": "key_required", "hint": ...} — that source needs an API key the user has not configured. Recovery: proceed with the hits you got (other sources answered); mention the hint's env var only if the user asks for that specific source's coverage.
  • {"source": ..., "reason": "error", "detail": ...} — the source's API failed (timeout, 429, 5xx). Semantic Scholar 429s on the shared pool without a key — this is normal, not a bug. Recovery: use the remaining hits; retry later only if that source is essential to the question.

The call itself raises (protocol isError: true) only for bad arguments:

  • sort must be one of: relevance, citations, date — fix the sort value.
  • unknown source(s): ... Known: ... — fix the sources list to names from list_sources.

hits: [] with nothing skipped means the query genuinely matched nothing: broaden keywords; do not invent results.

get_paper

Error-shaped results are dicts with an error key:

  • unknown id_type '...' — fix id_type to one of: doi, arxiv, pmid, openalex, s2 (or use "auto").
  • paper not found via <Source> (id_type=..., identifier=...) — the owning source has no record; the message carries its own numbered recovery steps. Retrying the same identifier will not help: check it for typos; if it was guessed by id_type: auto, pass the correct id_type explicitly; or search_papers for the title instead.

Successful results can still carry degraded legs, reported in notes:

  • references unavailable: ... / citations unavailable: ... — that leg's API failed; the paper itself is valid.
  • <Source> does not expose a public reference list — arXiv and PubMed have no public reference graph; this is a source limit, not a failure.
  • citing works not available from this source (OpenAlex fallback applies when the work is indexed there) — empty citations may mean "not indexed", never "never cited".

verification (get_paper verify=true)

  • resolves: true — landing page answers; false — 404/410, likely dead identifier; null — target offline or bot-blocked: say "could not verify", never claim the paper is dead.
  • retracted: true — never present the paper as valid evidence. The result additionally carries a RELAY TO THE USER text block (annotated audience: ["user"]): pass it on to the human verbatim.
  • retracted: null — retraction status unknown (OpenAlex does not index it); absence of a flag is not proof of validity.

skills

  • Skill '...' not found — call skills_list and use an exact name.
  • temporarily unavailable — fetch failed; proceed with tool docstrings and get_research_method.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.