datagma

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

• Third-party content exposure detected (high risk: 0.80). The skill's Popular actions and workflow instruct using Membrane to run Datagma actions (e.g., "Get Twitter Profile by Email", "Get Twitter Profile by Username", "Enrich Person", "Enrich Company" from LinkedIn URLs and news) which clearly fetch public, user-generated third‑party profiles/pages via the Datagma API and return that content for the agent to read and act on (see the "Popular actions" and "Running actions" sections in SKILL.md), so untrusted external content can influence decisions.