Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

biomedical-paper-billing

v1.0.0

AI-powered biomedical manuscript generation with docx output. Activates when user provides Chinese draft/outline and requests full English research paper. In...

0· 51·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
medium confidence
!
Purpose & Capability
The SKILL.md and assets/docx_builder.py consistently implement a biomedical manuscript -> DOCX generator, which matches the stated purpose. However, repository metadata is inconsistent: _meta.json displayName includes 'Billing/Enhanced' while the public registry listing name/description do not describe any billing/payment functionality; package.json claims a Node main (index.js) that is not present (the package is Python-based). These mismatches (owner/displayName differences and missing index.js) are unexplained and indicate sloppy packaging or possible repackaging of unrelated components.
Instruction Scope
The runtime instructions are explicit and focused: accept Chinese draft, extract text and references, validate every reference against PubMed (or batch_web_search), replace NOT_FOUND entries with recent high-quality papers, and produce numbered Vancouver references and a mapping table. Network activity (PubMed/web searches) is required by the workflow. The instructions also authorize automatic replacement of user-provided citations if they cannot be verified; while coherent with the declared goal (produce a 'clean' reference list), this behavior can materially change meaning and provenance of the user's draft and raises academic-integrity risk. The SKILL.md does not instruct reading any unrelated system files or environment variables.
Install Mechanism
No install spec is present (instruction-only skill) and the included Python file is a straightforward docx builder. Nothing is downloaded from untrusted URLs. Note: the code assumes python-docx is available but the skill provides no installer to ensure dependencies are present.
Credentials
The skill requests no environment variables, credentials, or config paths. There are no declarations requiring external secrets. One minor packaging inconsistency: package.json is for a Node package (index.js) even though the runtime is Python; this does not affect credentials but is an unexplained mismatch.
Persistence & Privilege
The skill does not request always:true and uses normal invocation semantics. It does not request system-wide configuration changes or persistent privileges in the files provided.
What to consider before installing
Things to consider before installing/using: - Metadata mismatches: _meta.json lists a different displayName/owner and package.json references index.js (missing). This could indicate careless packaging or repackaging — verify the author identity before trusting outputs. - No credentials required, but the SKILL.md instructs the agent to run PubMed/web searches and to automatically replace any user-supplied references that cannot be verified. That replacement behavior can change the provenance and claims in a manuscript — always manually review the 'citation mapping table' and substituted references. - The included Python file (assets/docx_builder.py) appears benign and only builds .docx files, but the runtime requires python-docx; the skill does not install dependencies for you. Run it in a sandbox or environment you control and ensure dependencies are pinned/inspected. - Ethical risks: this skill can facilitate producing academic manuscripts and could be used to generate misleading or inappropriate papers. Do not provide identifiable or unpublished patient-level data, and verify all statistics and references yourself before submission. - Practical steps: (1) confirm the registry owner and author identity, (2) test the tool on non-sensitive drafts, (3) inspect any network calls your agent will make (which PubMed endpoints or web_search tools are used), (4) review the final document and the '引用编号映射表' for any substituted citations, and (5) if you need a stricter audit, request a version with clear provenance for all web lookups (e.g., PMIDs and retrieval timestamps) and a manifest explaining why package.json/index.js exist.

Like a lobster shell, security has layers — review code before you run it.

latestvk97daa03zgbwpbe3mdd7z4t90n83q7kz

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

Comments