reply-api

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs the agent to fetch runtime docs (e.g. https://docs.reply.io/llms.txt and per-page markdown like https://docs.reply.io/api-reference/sequences/list-all-sequences.md) which are retrieved each session to determine exact endpoints and request bodies, so external content controls agent instructions.