nemoclaw-maintainer-verify-stale

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

• Third-party content exposure detected (high risk: 0.85). The required runtime workflow ingests outsider-authored free text from GitHub issue bodies/comments (e.g., Step 1/3/4 fetch body and comments via gh issue view --json ... and Step 6/8 use that text to extract/synthesize reproducer.sh, which is then executed and also fed into the agent’s LLM context for synth-repro/comment drafting), creating an indirect prompt-injection surface from outsider content.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill fetches and executes remote installers at runtime (notably curl -fsSL $INSTALL_URL | bash with INSTALL_URL defaulting to https://www.nvidia.com/nemoclaw.sh, plus runtime installs like curl -fsSL https://ollama.com/install.sh | sh and downloads from GitHub releases such as https://github.com/NVIDIA/OpenShell/releases/download/...), so external content is fetched during execution and run as code and is a required dependency.