lfg

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (medium risk: 0.30). The skill is a legitimate autonomous pipeline, but the verification section tells the agent that invoking lfg counts as user approval that overrides project instructions requiring it to ask first, which weakens a user-set protection beyond what the advertised task needs without malicious intent.

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

• Third-party content exposure detected (medium risk: 0.30). The required intake step reads a tracker ticket (GitHub issue, Jira, or Linear) named by an issue reference, and that ticket's free text, including its title, is outsider-authored and feeds the defect and planning routes, which counts as medium under the ticket-reading rule.