Security checks across malware telemetry and agentic risk
ClawSergeant is a coherent OpenClaw training skill, but it sends training content to external LLM/agent services and persists training memory with limited runtime control.
Install only for a controlled OpenClaw training setup. Confirm the target CLAW_RECIPIENT, avoid confidential prompts or agent outputs, use a limited API key, inspect or disable lesson logging and MEMORY.md writes if persistent memory is not intended, and obtain/review the missing learning_logger.py before running.
async def chat(
self, conversation: Conversation, temperature: float = 0.7
) -> str:
"""Send the conversation to the LLM and return the assistant reply."""
if not self._client:
raise RuntimeError("LLMHandler not started; call start() first")async def chat_json(
self, conversation: Conversation, temperature: float = 0.3
) -> dict:
"""Send the conversation and parse the reply as JSON.
Uses the response_format API parameter to enforce valid JSON output.
"""httpx>=0.27 loguru>=0.7 python-dotenv>=1.0
httpx>=0.27 loguru>=0.7 python-dotenv>=1.0
httpx>=0.27 loguru>=0.7 python-dotenv>=1.0
66/66 vendors flagged this skill as clean.