Skip to content
skillsafebeta
← Report library
Purpose / Other

Firecrawl Research Index Skill Security Audit

What the author says it does (original text)

Find the papers that answer a research query in Firecrawl's research paper index — a corpus of paper abstracts whose largest share is biomedical and life-science literature (PubMed, bioRxiv, medRxiv), alongside arXiv preprints in CS, physics, and math — using semantic search, semantic and structural expansion, and in-body verification. Use this skill for literature-finding and paper-retrieval task

Independent security check

Security risks found

Files checked
1
Risks found
2
Could it run dangerous commands?Looks for programs run straight after downloading, remote control of your computer, and hidden commands.No risks found
Could it expose your files or keys?Looks for uploads of files containing passwords or keys, and keys written directly in the code.Risks found: 1
Medium risk

Research queries and paper-verification questions are sent to Firecrawl

Source references: 4
What we found

The search command directly accepts query text, and the paper-reading command additionally accepts a specific question. Because the Skill expressly covers clinical, drug, gene, and disease topics, personal medical details, confidential research directions, drug codenames, or other private context placed in those fields would enter an external tool call.

Why this matters

The external service may receive and log sensitive health interests, research intent, or commercial R&D information. The provided file does not state how that information is retained, accessed, or reused.

The Skill directs the agent to submit the user's query and paper-specific verification questions through Firecrawl MCP tools or its CLI. Because its scope includes clinical, drug, gene, and disease topics, personal health details, unpublished research, or internal project names could be sent to an external service if included. The source does not state retention or privacy terms. Users can ask for Firecrawl's data-handling policy and restrict calls to de-identified, non-confidential search terms.

SKILL.md:12In the instructionsOpen original file
Paper abstracts, with full text reachable per paper. The largest share of the corpus is **biomedical and life-science** literature — **PubMed** journal articles plus **bioRxiv** and **medRxiv** preprints — so clinical, drug, gene, disease, epidemiology, and public-health questions are in scope. **arXiv** preprints cover computer science, physics, and mathematics. Coverage outside those sources is thinner: a paper that exists only behind a publisher paywall or in a niche venue may not be indexed, and the general web tools below are the fallback when it isn't.
Show 3 other places
SKILL.md:18In the instructionsOpen original file
- MCP: **`firecrawl_research_search_papers(query, k?)`**  CLI: **`firecrawl research search-papers <query> [--k <number>]`**  Semantic (HyDE) search over **abstracts**. The natural first move for almost any query.  If results look thin or all-alike, re-run with a different framing (sibling domain, rival method, dataset/benchmark name) rather than giving up.
SKILL.md:23In the instructionsOpen original file
- MCP: **`firecrawl_research_related_papers(seed_ids, intent, mode?, k?)`**  CLI: **`firecrawl research related-papers <seedIds...> --intent <intent> [--mode <similar|citers|references>] [--k <number>]`**  Semantic and structural expansion, ranked to your `intent`.
SKILL.md:35In the instructionsOpen original file
- MCP: **`firecrawl_research_read_paper(id, question)`**  CLI: **`firecrawl research read-paper <id> --question <question>`**  In-body passages of **one** paper, to verify a load-bearing constraint (a method actually used, a score actually reported, an affiliation, what a paper compares to).  Use it to settle a specific doubt, not on everything.
Could it delete files or keep running?Looks for broad file deletion, disk overwrites, and programs set to start automatically.No risks found
Could it bypass safety checks?Looks for skipped website security checks, excessive file access, or actions that skip your approval.No risks found
Could it mislead the AI or hide text?Checks the skill instructions for requests to ignore you, influence the report, or hide text in invisible characters.Risks found: 1
Medium risk

The “include when in doubt” policy can let weakly supported papers influence decisions

Source references: 3
What we found

The Skill tells the agent to retain most “plausibly relevant” papers and treats body verification mainly as a way to exclude them, rather than requiring candidates to be substantiated first. The index also contains bioRxiv, medRxiv, and arXiv preprints, while the instructions require no distinction for peer-review status, study quality, retraction, or risk of bias.

Why this matters

In high-impact clinical, drug, or public-health use, the list may include papers that are only topically similar, unverified, or methodologically weak. Treating inclusion as evidentiary support could lead to unreliable medical, research, or commercial judgments.

This goes beyond a recall-oriented search suggestion: the instructions explicitly retain most “plausibly relevant” papers and use full-text checks mainly to exclude candidates. The corpus includes bioRxiv, medRxiv, and arXiv preprints, while the visible instructions do not require peer-review, retraction, or study-quality labels. In medical or other high-stakes decisions, weakly relevant or unreviewed work could therefore appear alongside stronger evidence and influence judgment. Users can require evidence-level and publication-status labels, stated verification limits, and stricter inclusion criteria for decision use.

SKILL.md:12In the instructionsOpen original file
Paper abstracts, with full text reachable per paper. The largest share of the corpus is **biomedical and life-science** literature — **PubMed** journal articles plus **bioRxiv** and **medRxiv** preprints — so clinical, drug, gene, disease, epidemiology, and public-health questions are in scope. **arXiv** preprints cover computer science, physics, and mathematics. Coverage outside those sources is thinner: a paper that exists only behind a publisher paywall or in a niche venue may not be indexed, and the general web tools below are the fallback when it isn't.
Show 2 other places
SKILL.md:66In the instructionsOpen original file
- **Query shape and subject field are separate.** A clinical-trial question and a machine-learning question take the same shapes above; what differs is only which source the hits come from. Don't send a biomedical or life-science query to the open web on the assumption the corpus is arXiv-only — PubMed, bioRxiv, and medRxiv are the largest part of what `search_papers` reads.- **When in doubt, include.** For any topic / method / comparison question, return the relevant _family_, not just the single best match — err toward keeping a plausibly-relevant paper rather than dropping it. The neighboring methods are part of a good answer; don't reason close work out just because one paper is the most exact match.- **Follow the literature, and keep what you find.** The seminal source, the competing methods, the close neighbors are usually a hop away — use `related_papers`, and _include_ them, not just the first hit. Stopping at one good result is the most common way to leave the reader with half an answer.
SKILL.md:68In the instructionsOpen original file
- **Follow the literature, and keep what you find.** The seminal source, the competing methods, the close neighbors are usually a hop away — use `related_papers`, and _include_ them, not just the first hit. Stopping at one good result is the most common way to leave the reader with half an answer.- **Verify to exclude, not to gatekeep.** Use `read_paper` to rule a paper _out_ when a hard constraint clearly fails (wrong org/author, doesn't actually report the score). When a paper is plausibly relevant, lean toward keeping it rather than demanding proof.- **Only drop the clearly off-topic.** Don't pad with papers you're confident are unrelated — but that's a high bar; most plausibly-relevant work should make the cut.
Could it change links or payment recipients without asking?Looks for forced referral or payment changes combined with instructions to hide the change.No risks found

Inside this skill

5 instruction sections

The Skill sends research queries through Firecrawl MCP tools or CLI for semantic search over paper abstracts, then can expand results through similar papers, citers, and references.

View source
SKILL.md:18In the instructionsOpen original file
- MCP: **`firecrawl_research_search_papers(query, k?)`**  CLI: **`firecrawl research search-papers <query> [--k <number>]`**  Semantic (HyDE) search over **abstracts**. The natural first move for almost any query.  If results look thin or all-alike, re-run with a different framing (sibling domain, rival method, dataset/benchmark name) rather than giving up.
SKILL.md:23In the instructionsOpen original file
- MCP: **`firecrawl_research_related_papers(seed_ids, intent, mode?, k?)`**  CLI: **`firecrawl research related-papers <seedIds...> --intent <intent> [--mode <similar|citers|references>] [--k <number>]`**  Semantic and structural expansion, ranked to your `intent`.  This reaches papers semantic search _cannot_, and it's how you turn one good hit into the rest of a set.  `mode=similar` → niche siblings; `citers` → who uses/builds on the seeds; `references` → what they build on / compare against.

For questions requiring verification inside a paper, the Skill supplies a paper ID and a specific question to `read_paper`, which returns relevant body passages; it advises using this only to resolve a specific doubt.

View source
SKILL.md:35In the instructionsOpen original file
- MCP: **`firecrawl_research_read_paper(id, question)`**  CLI: **`firecrawl research read-paper <id> --question <question>`**  In-body passages of **one** paper, to verify a load-bearing constraint (a method actually used, a score actually reported, an affiliation, what a paper compares to).  Use it to settle a specific doubt, not on everything.

For leaderboard, largest, or most-popular questions, the Skill switches to ordinary web search or scraping and then maps ranked entries back to papers; this path is not confined to the paper index.

View source
SKILL.md:46In the instructionsOpen original file
- MCP: **`firecrawl_search(query)` / `firecrawl_scrape(url)`**  CLI: **`firecrawl search <query>` / `firecrawl scrape <url>`**  General **web** search and page fetch, for facts that don't live in paper abstracts: benchmark **leaderboards**, rankings, "who scores best / is largest / is most used."  Find the ranking on the web, then map the top entries back to papers with `search_papers`.  Reach for these only when the corpus can't answer the question on its own.

The Skill's selection policy favors returning a fuller set of potentially relevant papers, excluding candidates mainly when they are clearly off-topic or fail a hard constraint.

View source
SKILL.md:66In the instructionsOpen original file
- **Query shape and subject field are separate.** A clinical-trial question and a machine-learning question take the same shapes above; what differs is only which source the hits come from. Don't send a biomedical or life-science query to the open web on the assumption the corpus is arXiv-only — PubMed, bioRxiv, and medRxiv are the largest part of what `search_papers` reads.- **When in doubt, include.** For any topic / method / comparison question, return the relevant _family_, not just the single best match — err toward keeping a plausibly-relevant paper rather than dropping it. The neighboring methods are part of a good answer; don't reason close work out just because one paper is the most exact match.- **Follow the literature, and keep what you find.** The seminal source, the competing methods, the close neighbors are usually a hop away — use `related_papers`, and _include_ them, not just the first hit. Stopping at one good result is the most common way to leave the reader with half an answer.- **Verify to exclude, not to gatekeep.** Use `read_paper` to rule a paper _out_ when a hard constraint clearly fails (wrong org/author, doesn't actually report the score). When a paper is plausibly relevant, lean toward keeping it rather than demanding proof.- **Only drop the clearly off-topic.** Don't pad with papers you're confident are unrelated — but that's a high bar; most plausibly-relevant work should make the cut.
Start here · InstructionsSKILL.md
firecrawl-research-index
Lines connect the instruction file to its sections, not an observed execution order. Select a section to read the source.
Files and check records1 files

Coverage and gaps

Content covered in each file

These are the source ranges included in this check, not a guarantee that every issue has been resolved.

  • SKILL.mdFull text included

This report is for the version above. We read the available code and instructions without running the skill or checking extra packages it installs. This is not a promise of safety: a different version or setup may behave differently.

  • SKILL.mdInstructions

Operations mentioned in code and instructions

Connect to websites
SKILL.md:73In the instructionsOpen original file
- [firecrawl-build-search](https://github.com/firecrawl/skills/tree/main/skills/build/firecrawl-build-search) — building the paper index into an app instead of querying it here
Lines read
74
File checksum (to compare versions)
7fc95ca9c37654337d315c53f2aad3ea66892b5f4615e5c2f9c3db8a07eec104