r2http

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill instructs the agent to request user credentials and embed them verbatim in curl commands (e.g., curl -u user:pass ...), which requires the LLM to handle and output secret values directly.

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

• Third-party content exposure detected (medium risk: 0.30). SKILL.md describes using user-supplied R2CMD/endpoint data to send raw radare2 commands via POST $R2CMD with curl, so the required runtime workflow can ingest outsider-authored free text as the HTTP request body without selecting a specific trusted item.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The skill instructs the agent to start/modify a radare2 HTTP server (including disabling the sandbox with http.sandbox=false and optionally binding to public interfaces), and to run background/long-lived processes — actions that change system and network state and can bypass local security protections.