中文 EN

Scan Report

Scan New Files

⚠ Low Risk — Risk Score 20/100
Last scan:20 hr ago Rescan
20 /100
format-flow
Multi-format document conversion toolkit supporting Word ↔ PDF ↔ Markdown, web pages, text formatting, Excel to JSON, and image processing
A legitimate document conversion toolkit with minor undocumented behaviors (auto-pip install, LibreOffice subprocess, network requests) that are necessary for core functionality.
Skill Nameformat-flow
Duration52.4s
Enginepi
✓
Safe to install
Consider adding explicit documentation about subprocess usage for dependency installation and network access for web conversion features. Pin dependency versions for supply chain safety.

Findings 3 items

Severity Finding Location
Low
Undeclared network access Doc Mismatch
The web_to_markdown.py module makes HTTP GET requests to URLs, but SKILL.md does not declare network:READ capability
response = requests.get(url, headers=headers, timeout=30)
→ Add network:READ to allowed-tools mapping in skill metadata or document network usage in SKILL.md
 scripts/converters/web_to_markdown.py:89
Low
Unversioned dependency installation Supply Chain
Dependencies are installed via pip without version pinning, creating supply chain risk
subprocess.check_call([sys.executable, '-m', 'pip', 'install', package])
→ Use version-pinned requirements or lock files: pip install package==1.2.3
 scripts/utils/dependencies.py:57
Low
Undeclared subprocess usage RCE
LibreOffice 'soffice' command executed via subprocess for PDF conversion, not declared in documentation
subprocess.run(cmd, check=True, capture_output=True)
→ Document shell:WRITE capability or provide alternative pure-Python fallback
 scripts/converters/word_to_pdf.py:54
ResourceDeclaredInferredStatusEvidence
Filesystem WRITE WRITE ✓ Aligned All converter modules read/write user-specified files
Network NONE READ ✓ Aligned web_to_markdown.py:89 requests.get(url)
Shell NONE WRITE ✓ Aligned dependencies.py:57 subprocess.check_call; word_to_pdf.py:54 soffice
Environment NONE NONE — No os.environ access found
Skill Invoke NONE NONE — No skill self-invocation
Clipboard NONE NONE — No clipboard access
Browser NONE NONE — No browser automation
Database NONE NONE — No database access
1 findings
🔗
Medium External URL 外部 URL
https://pandoc.org/installing.html
README.md:39

File Tree

15 files · 115.6 KB · 3903 lines
Python 13f · 3140L Markdown 2f · 763L
├─ ▾ 📁 scripts
│ ├─ ▾ 📁 converters
│ │ ├─ 🐍 __init__.py Python 87L · 1.8 KB
│ │ ├─ 🐍 excel_to_json.py Python 318L · 9.5 KB
│ │ ├─ 🐍 image_processor.py Python 417L · 12.5 KB
│ │ ├─ 🐍 markdown_to_word.py Python 182L · 5.4 KB
│ │ ├─ 🐍 pdf_to_markdown.py Python 187L · 5.0 KB
│ │ ├─ 🐍 text_formatter.py Python 406L · 11.1 KB
│ │ ├─ 🐍 web_to_markdown.py Python 385L · 11.6 KB
│ │ ├─ 🐍 word_to_markdown.py Python 273L · 7.8 KB
│ │ └─ 🐍 word_to_pdf.py Python 131L · 3.8 KB
│ ├─ ▾ 📁 utils
│ │ ├─ 🐍 __init__.py Python 46L · 1.0 KB
│ │ ├─ 🐍 dependencies.py Python 191L · 5.6 KB
│ │ └─ 🐍 helpers.py Python 110L · 2.7 KB
│ └─ 🐍 convert.py Python 407L · 17.1 KB
├─ 📝 README.md Markdown 337L · 7.7 KB
└─ 📝 SKILL.md Markdown 426L · 13.1 KB

Dependencies 7 items

PackageVersionSourceKnown VulnsNotes
python-docx * pip No Version not pinned
pdfplumber * pip No Version not pinned
Pillow * pip No Version not pinned
requests * pip No Version not pinned
beautifulsoup4 * pip No Version not pinned
openpyxl * pip No Version not pinned
tqdm * pip No Version not pinned

Security Positives

✓ No credential harvesting - no access to API keys, tokens, or passwords
✓ No data exfiltration - no external data transmission
✓ No obfuscation - clean, readable code without base64 or eval
✓ No sensitive path access - no ~/.ssh, ~/.aws, or .env file reading
✓ No persistence mechanisms - no cron jobs or startup hooks
✓ No reverse shell or C2 infrastructure
✓ All file operations are user-controlled (input/output paths specified by user)