k-skill-cleaner

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The workflow reads local user chat logs and JSONL session files (~/.claude/projects/**/*.jsonl, etc.) which can contain untrusted user- or outsider-authored text injected into session transcripts.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill fetches and installs remote content via npx -y @nomadamas/k-skill@0, where @nomadamas is an individual-owned GitHub repository/NPM scope lacking official organizational provenance, establishing a high-tier unverifiable runtime external dependency.