Security risk report for @forgecat/anthropics_skills_pdf v0.0.5
Source Integrity
Low
Profile claims origin from Anthropic's official skills repository (https://github.com/anthropics/skills) with commit hash 5128e18.
Metadata indicates legitimate source platform (Claude Code) and standard license (LicenseRef-Anthropic-Terms).
Static scanner confirms low risk; no supply-chain red flags detected.
Agent Intent
Low
Content is purely educational documentation of legitimate PDF processing libraries (pypdf, pdfplumber, reportlab, pdf-lib, pdfjs-dist) with standard usage examples.
No instructions to ignore system prompts, exfiltrate data, read credentials, or execute remote payloads; no role-hijacking or prompt-injection language.
Guidance on PDF form filling (FORMS.md) and structure extraction is procedural and security-neutral; no poisoning of dependencies or weakening of security defaults.
Details
Evidence
README.md: 'Use this skill whenever the user wants to do anything with PDF files. This includes reading or extracting text/tables from PDFs...'
SKILL.md: Standard Python library examples (pypdf, pdfplumber, reportlab) with no malicious instructions.
FORMS.md: Step-by-step procedural guidance for filling PDF forms using legitimate tools; no hidden exfiltration or credential theft.
REFERENCE.md: Advanced library documentation (pypdfium2, pdf-lib, pdfjs-dist) with standard code patterns; no obfuscated payloads.
Permissions
Low
Profile declares no tools or MCP servers; it is purely instructional documentation for PDF processing.
No file_write, file_delete, shell, or other high-risk tool categories are invoked or configured.
Authority is implicit to the stated purpose: guiding users on PDF manipulation using standard libraries.
Details
Evidence
Declared tools section: (none declared)
MCP servers section: (none)
Content is markdown documentation only; no tool definitions or permission grants.
MCP Risk
Low
No MCP servers are defined in the profile.
Content relies on standard, well-known Python and JavaScript libraries available via public package managers (pip, npm).
No hidden instructions in tool descriptions, arbitrary binary execution, or unrestricted network/filesystem access.
Details
Evidence
MCP servers section: (none)
Libraries referenced are all public and auditable: pypdf, pdfplumber, reportlab, pytesseract, pdf2image, pypdfium2, pdf-lib, pdfjs-dist, poppler-utils, qpdf, pdftk.
ForgeCat