gangprompting

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

• Third-party content exposure detected (high risk: 0.95). Yes: the required runtime workflow runs discord-agent-bridge.ts monitor and pipes its NDJSON (containing content from Discord users) into opencode-relay.ts, which injects it into the OpenCode session via POST /session/:id/prompt_async as free text (parts[0].text), so outsider-authored Discord message content can reach the agent’s LLM context.