nodetool

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The prompt includes examples that pass auth tokens/secrets directly in command-line arguments and stdin JSON (e.g., --auth-token "my_auth_token", echo '{"auth_token":"token"}' | ...), which would require the agent to embed secret values verbatim into generated commands/outputs and thus poses high exfiltration risk.

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.90). Both URLs point to raw install scripts (.sh and .ps1) intended to be piped to a shell/PowerShell — a pattern that can execute arbitrary commands from the remote repo (even if hosted on raw.githubusercontent.com), so unless you verify the repository/authenticity and inspect the scripts first, this is high risk.

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

• Third-party content exposure detected (high risk: 0.70). The skill's documentation and CLI include explicit runtime commands that fetch and load third‑party, user‑hosted artifacts—e.g., "nodetool admin download-hf <model_name>" and "nodetool admin download-ollama" for HuggingFace/Ollama models and the install instructions that curl a script from raw.githubusercontent.com—so untrusted external content can be ingested and materially influence model behavior.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's install instructions fetch and immediately execute remote installer scripts (e.g., curl -fsSL https://raw.githubusercontent.com/nodetool-ai/nodetool/refs/heads/main/install.sh | bash and irm https://raw.githubusercontent.com/nodetool-ai/nodetool/refs/heads/main/install.ps1 | iex), which runs remote code at runtime and are required for setup.