writing-social-posts

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

• Third-party content exposure detected (low risk: 0.10). In the writing-social-posts workflow, step 6 “Postflight” and step 1 “Find the exemplar post first” can read outsider-authored free text from user-supplied URLs/status URLs (e.g., via curl ... https://cdn.syndication.twimg.com/tweet-result?... and by reading “the live copy” at the live URL).

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

• Potentially malicious external URL detected (high risk: 0.70). The skill runs a runtime curl to "https://cdn.syndication.twimg.com/tweet-result?id=<POST_ID>&token=a" (SKILL.md) to fetch a remote post which is then read and used as the exemplar shape for drafting, so external content is injected into the agent's context and can directly control output.