recoup-content-write-caption

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

• Third-party content exposure detected (medium risk: 0.65). Outsider-authored free text can enter the LLM context via the “fallback to 10–30 real past captions” sourced from /artists/{ROW_ID}/posts (API-fetched post captions authored by others, then read/used to draft captions).

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

• Potentially malicious external URL detected (high risk: 0.80). The skill contains runtime curl calls to Recoup APIs (e.g. "https://api.recoupable.dev/api" and "https://api.recoupable.dev/api/content/analyze") whose responses (research signals, templates, analyze text, artist posts) are fetched at runtime and are used to steer prompt/template selection and generation, so they directly control agent instructions.