fluent-2

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

• Third-party content exposure detected (high risk: 0.80). SKILL.md step 3-4 instructs running curl to fetch live catalog text (https://api.agent-ds.oday-bakkour.com/.../design.md) and step 2 says to copy/merge it, so that fetched outsider-authored free-form markdown can be incorporated into the agent’s LLM context.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill includes an explicit runtime fetch command that can replace or refresh the agent's design/prompt instructions by downloading https://api.agent-ds.oday-bakkour.com/v1/systems/fluent-2/design.md, which means remote content can directly control agent prompts.