lovstudio-maintain-partners

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly fetches and renders public third‑party webpages and assets (see SKILL.md Op 1 step 2 and scripts/scrape_logo.py which curl/Playwright‑fetch brand homepages, and scripts/audit_partners.py --probe which HTTP‑probes hrefs), and those untrusted remote pages/images are read and used by the workflow (downloaded logo candidates, visual verification, normalization, and decisions like --show-name/--invert), so external content can materially influence the agent's actions.