openclaw-telegram-bot-debug

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

• Third-party content exposure detected (high risk: 0.70). The SKILL.md explicitly instructs inspecting Telegram channel logs with "openclaw channels logs --channel telegram --lines 200" (and to read gateway logs/probe channel status), which causes the agent/workflow to ingest untrusted, user‑generated content from Telegram and to use those log contents to drive diagnostic decisions.

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

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The playbook instructs modifying systemd user service files, copying/editing root-owned config and token files, and running privileged commands like loginctl enable-linger and sudo -u ..., which require elevated privileges and change system state.