epic-captain

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The skill deliberately ships repository worktrees and local env files to external LLM executors (codex/claude) with workspace-write/--add-dir and --dangerously-skip-permissions, enabling exfiltration of repo data and remote-driven commits/pushes that could be abused to inject backdoors or leak secrets.

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

• Third-party content exposure detected (high risk: 0.85). Yes: scripts/implement-issue.sh fetches Linear issue description/comments (outsider-authored tracker content) and writes them into $WT/.impl-spec.md, which is then embedded verbatim into the executor LLM prompt via codex exec ... "$PROMPT".