opencode

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.85). This skill intentionally instructs use of auto-approve flags and permissive environment/config settings (e.g., --dangerously-skip-permissions, OPENCODE_DANGEROUSLY_SKIP_PERMISSIONS, OPENCODE_CONFIG_CONTENT='{"permission":{"*":"allow"}}'), plus agent modes that grant filesystem and shell access and a persistent server mode — behaviors that enable headless remote code execution, unattended file modifications, and could be abused for backdoors or data exfiltration.