prisma

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

• Third-party content exposure detected (high risk: 0.85). The required workflow instructs the agent to run WebFetch("https://www.prisma.io/docs/llms.txt"), which is public web content fetched at runtime and thus outsider-authored free text that can be injected into the LLM context.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly requires a runtime fetch WebFetch("https://www.prisma.io/docs/llms.txt") before proceeding, so external content from that URL is a required dependency that directly controls the agent's instructions.