autonomous-loops

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.80). The link is a direct raw GitHub URL to an install.sh script (commonly used with "curl … | bash"); executing unreviewed remote shell scripts is high risk because they can run arbitrary commands even if hosted on GitHub and should only be used after verifying the repo, file contents, and integrity.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The skill instructs automated agents to send full repo/session context to external LLMs, run arbitrary shell tools (including piping a remote install script), and automatically commit/push/merge changes — patterns that enable data exfiltration, remote code execution, credential harvesting, and supply‑chain compromise if abused.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill includes an installation command that fetches and pipes a remote shell script to bash—executing remote code at runtime—via https://raw.githubusercontent.com/AnandChowdhary/continuous-claude/HEAD/install.sh, which is a required external dependency for the described "continuous-claude" workflow.