genviral

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs the agent to fetch and import public TikTok content (copy-tiktok-preview and copy-tiktok-import) and to browse TikTok search pages with the browser tool during competitor research (docs/references/competitor-research.md), so untrusted, user-generated third-party content is read and used to drive generation and posting decisions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill includes a runtime self-updater (scripts/update-skill.sh) that fetches and applies updates from the GitHub repo https://github.com/fdarkaou/genviral-skill, which can modify SKILL.md and other agent files that directly control prompts/instructions — this is a high-confidence runtime fetch that can change agent behavior.