task

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

• Third-party content exposure detected (high risk: 0.90). This skill's intake and tracked-task rules require fetching and reading third-party, user-generated tracker content (e.g., GitHub issues/PRs via gh issue view/gh pr view, GitHub/Linear comments and attachments, and tracker-hosted video transcripts) as the source-of-truth, which allows untrusted content to influence agent decisions and actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly fetches GitHub issue/PR content at runtime using gh issue view / gh pr view (i.e. https://github.com/... issue/PR URLs), and treats that fetched text as the source-of-truth that directly drives agent instructions, so remote content can control prompts.