skill-finder

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

• Third-party content exposure detected (medium risk: 0.30). The workflow queries a community-driven skills registry (claude-plugins.dev) and automatically installs or fetches external skills created by third-party authors, exposing the agent to outsider-authored free text in skill descriptions and downloadable SKILL.md files.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill queries an untrusted third-party community plugin registry (claude-plugins.dev) and executes arbitrary third-party code packages downloaded via npx or manual raw file installation, introducing a high-risk unverifiable runtime external dependency.