logic-master

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The skill deliberately mandates an "auto-approve" --yolo mode that bypasses sandbox/safety and forces operators to run CLI commands with elevated file-system and execution privileges, which is an intentional backdoor-like design that enables remote code execution and abuse (high-risk for misuse and data exfiltration).

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

• Third-party content exposure detected (medium risk: 0.30). The Codex Logic Skill runtime workflow ingests outsider-authored free text as the direct instruction passed to codex exec --yolo (and optionally via stdin/file/image paths), without requiring selection of a specific trusted item first.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.90). The skill repeatedly mandates use of a "--yolo" flag that "auto-approve[s] and bypass[es] sandbox" and instructs executing bash/Edit tools and applying changes, which encourages breaking sandbox protections and allows arbitrary modifications to the host (including privileged or persistent changes) even if it doesn't explicitly mention sudo or user creation.