file-security-toolkit
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted document content, creating a surface where malicious instructions could influence agent behavior.
- Ingestion points: The
redactfunctionality inscripts/file_security.pyreads data from user-provided txt, md, csv, docx, and pptx files. - Boundary markers: Absent. The skill does not utilize delimiters or specific instructions to warn the agent about embedded content in processed files.
- Capability inventory: The skill possesses file system read/write capabilities through its implementation script.
- Sanitization: The script targets specific PII patterns for redaction (emails, phone numbers, IDs) but does not filter for potential prompt injection instructions.
- Self-referential content: The file
file-security-toolkit_audit_result_v1.jsoncontains self-authored safety claims and audit scores ("security": "PASS") designed to influence the security evaluation process. - [EXTERNAL_DOWNLOADS]: The skill documentation specifies several third-party Python libraries required for its core operations.
- Dependencies include
pyzipper,cryptography,python-docx,python-pptx, andpillowfor encryption and document handling. - References are made to the author's GitHub repository at
github.com/aipoch/medical-research-skillsfor source code and documentation.
Audit Metadata