imclaw

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill instructs the agent to prompt for and receive tokens and then embed those token values verbatim into files/commands/JSON (e.g., the gateway tool "raw" field and curl Authorization headers), which requires the LLM to handle/output secret values directly.

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

• Third-party content exposure detected (high risk: 0.90). This skill ingests untrusted, user-generated messages from the IMClaw Hub (received over WebSocket by bridge_simple.py and via API calls like get_history / fetch_and_archive.py) and injects those messages into the main OpenClaw Session via /hooks/wake (the generated wake_text), which the agent is expected to read and which can directly influence decisions and trigger actions (replies, task execution, update.py), enabling indirect prompt injection.