repo-scan

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

• Third-party content exposure detected (high risk: 0.85). The skill is a repo-scanning tool that, at runtime, enumerates and reads the target repository’s files (including third-party/vendored code and other contributors’ source text) and feeds extracted prose/metadata into the LLM context for analysis, which is outsider-authored content from the scanned repo.

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

• Potentially malicious external URL detected (high risk: 1.00). The installation steps fetch and install remote code from the Git URL https://github.com/haibindev/repo-scan.git into the agent's skills directory, which means runtime behavior and prompts could be controlled by that external repository's code.