Skill flagged — suspicious patterns detected

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

Clinical Case Writer

v2.0.1

根据患者病历资料撰写符合GB/T 7713.2-2022规范的临床案例报告,自动生成格式规范的Word文档。

0· 21·0 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
SKILL.md and README promise automatic extraction from PDF/Word/images and automatic literature retrieval (知网/CNKI + PubMed). The only included script (clinical_case_writer.py) implements Markdown→docx conversion and formatting/validation; it contains no OCR, PDF parsing, web retrieval, or CNKI/PubMed API calls. Declared features (automatic extraction, web retrieval and filtering, core-journal filtering) are therefore unmet by the code bundle.
!
Instruction Scope
Runtime instructions tell the agent to read patient records (PDF/Word/images), extract PHI and key information, and search CNKI/PubMed with specific filters (勾选'核心期刊', date ranges, exclude reviews/guidelines). Those actions imply file I/O, OCR/parsing, and web access, but the skill does not declare tools, credentials, or safe handling of protected health information. The SKILL.md also references memory/knowledge files and a generate script name that exists, but the referenced knowledge documents are not included in the package manifest. The instructions therefore grant broad discretion without describing the mechanisms or constraints.
Install Mechanism
No install spec; instruction-only with a small Python script. Requirements in skill.json list python-docx, which matches the script's use. No downloads, remote installers, or unusual install behavior present.
Credentials
The skill requests no environment variables, no credentials, and no config paths — this is proportionate for the included docx-generation script. However, SKILL.md expects the agent to perform online CNKI searches and possibly access external systems; that would normally require network access or credentials (particularly for CNKI), but none are declared. The absence of declared credentials is inconsistent with the claimed automatic retrieval capability.
Persistence & Privilege
always is false and autonomous invocation is allowed by default. The skill does not request persistent or elevated privileges and does not modify other skills' configurations. No concern on persistence from the package contents.
What to consider before installing
This skill's documentation promises automated extraction from patient records and automatic literature retrieval, but the only code provided converts Markdown into a formatted Word document and validates references. Before installing or using: 1) Do not assume the skill will perform PDF/OCR or web searches — those capabilities are not implemented and would require additional tools, network access, or credentials (especially for CNKI). 2) Protect patient privacy: the instructions reference reading patient files; ensure no PHI is uploaded to external search services and verify the agent will anonymize names as required. 3) Verify missing files: SKILL.md refers to knowledge documents that are not bundled — ask the author for the referenced memory/knowledge files or inspect how the agent will obtain them. 4) Test locally in a sandbox: run the Python script offline with sample Markdown to confirm behavior and check that it does not make network calls. 5) If you expect automated literature retrieval, require explicit explanations of how the skill will access CNKI/PubMed (APIs, scraping, credentials) and confirm compliance with site terms and data governance. If these gaps are acceptable and you will provide the missing extraction/retrieval components yourself, the docx-generation part appears benign; otherwise treat the skill as incomplete and verify origins/author before trusting with PHI or automating web searches.

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

latestvk975rjmd2kmgsmrm2p3z6ttq8n848db5

License

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

Comments