antigravity-swarm

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This skill directly executes and writes content produced by untrusted LLM output (<<RUN_COMMAND>> is executed with shell=True and <<WRITE_FILE>> can overwrite arbitrary files) and also sends workspace files and prompts to the external "gemini" CLI, enabling remote code execution and potential data exfiltration of local secrets—high-risk behavior.