repomix

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

• Third-party content exposure detected (high risk: 0.90). The skill's SKILL.md and usage patterns explicitly show it can fetch and package remote public repositories (e.g., "Remote Repository Support" and commands like bunx repomix --remote https://github.com/owner/repo and the "Third-Party Library" security-audit workflow), and then instruct the agent to read and analyze the packaged output, so untrusted public repo content can be ingested and influence analysis/actions.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill and its scripts invoke repomix at runtime with --remote to fetch GitHub repositories (e.g., https://github.com/owner/repo and the example https://github.com/yamadashy/repomix), and those fetched repository files are packaged for and fed into LLM context — meaning external content fetched at runtime can directly influence model prompts/behavior.