securityclaw-autonomous-soc-agent

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

• Third-party content exposure detected (medium risk: 0.30). The SecurityClaw SOC agent monitors and processes network logs and security events from OpenSearch databases using LLM-powered skills like threat_analyst and custom_detector.

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

• Potentially malicious external URL detected (high risk: 0.70). The installation instructions explicitly instruct the agent to clone an external GitHub repository (https://github.com/SecurityClaw/SecurityClaw.git) and execute installation commands (pip install -r requirements.txt), constituting an unverified third-party runtime dependency.