lossless-claw-skill

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly ingests arbitrary user conversation messages and stored files from OpenClaw sessions and exposes them to agents/sub-agents (see reference/architecture.md about persisted messages and XML-wrapped context, and reference/agent-tools.md / SKILL.md describing lcm_grep, lcm_describe returning full file content, and lcm_expand_query which spawns sub-agents that read source messages), so untrusted user-generated content can directly influence agent behavior.