promote-on-x

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly navigates and reads public X (Twitter) content (e.g., "Navigate to the feed source (home page)" and references/X_INTERFACE.md with navigate/read_page/open permalink instructions for https://x.com/home), ingests user-generated posts and replies to decide which tweets to target and what to post, so untrusted third‑party content directly influences agent actions.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly navigates to and reads live content from X (e.g., https://x.com/home and tweet permalinks like https://x.com/{username}/status/{tweet_id}) at runtime and injects those fetched posts into the agent's context to drive reply generation, so the external X URLs directly control the agent's prompts.