dreamer

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

• Third-party content exposure detected (high risk: 0.75). The skill’s runtime workflow uses gh issue list to fetch outsider-authored GitHub issue body text (notably from the factorium:graveyard issues) and includes that readable prose in the agent’s context for ideation/deduping.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill runs GitHub CLI commands at runtime (e.g., "gh issue list --repo {REPO} --label ... --json ..." and "gh issue create --repo {REPO} ..."), which fetch external GitHub issue content that is then incorporated into the agent's context/prompts, so remote GitHub data directly controls the agent's outputs.