ljg-skill-map
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill functions as a metadata viewer for local agent installations. It programmatically iterates through skill directories to extract frontmatter from
SKILL.mdfiles. This is a legitimate administrative capability with no evidence of data exfiltration or malicious intent. - [PROMPT_INJECTION]: The skill exposes a surface for indirect prompt injection by processing and rendering descriptions from other skills (Category 8).
- Ingestion points: Reads metadata from
SKILL.mdfiles located in~/.claude/skills/and~/.workbuddy/skills/. - Boundary markers: Output is constrained within defined ASCII box drawing characters to separate it from the main dialogue flow.
- Capability inventory: The skill performs local file reading and categorized printing to stdout; it lacks network access, file-writing capabilities, or arbitrary code execution tools.
- Sanitization: Both the Bash and Python components implement sanitization by removing newlines from descriptions, truncating text to the first sentence or 80 characters, and escaping quotes/backslashes to ensure the resulting output does not break formatting or execute as instructions.
Audit Metadata