gepetto

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The skill contains multiple high-risk patterns: it routinely sends full plan and (potentially) codebase content to external CLIs/LLMs, includes non-interactive flags that bypass safety checks, and generates prompts for autonomous execution (ralph-loop / Ralphy) that can modify repositories or run commands—behaviors that can be abused for data exfiltration and remote code execution.

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

• Third-party content exposure detected (medium risk: 0.30). Step 5 “Execute Research” launches web-search/explore subagents and can WebFetch/extract recommendations into claude-research.md, which exposes the agent to outsider-authored free text from the web.