add-opencode

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

• Third-party content exposure detected (high risk: 0.80). This skill wires the OpenCode agent provider into the runtime and explicitly directs the agent to call external provider APIs (see OPENCODE_PROVIDER/OPENCODE_MODEL and ANTHROPIC_BASE_URL entries such as https://openrouter.ai/api/v1 and https://api.deepseek.com/v1), whose untrusted model responses the agent consumes and can influence its subsequent actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly runs git fetch / git show against the repository "origin" (e.g., via "git fetch origin providers" and "git show origin/providers:...") to pull provider source files into the runtime image — that fetched remote code directly becomes the provider implementation that controls agent prompts/behavior and is a required dependency.