@dharma-ai-labs/agent-fabric@0.2.164
Local-first CLI for connecting coding agents to Dharma evaluation, orchestration, and signed Skill delivery
AI security review
scanned · by lpm-firewall-aiThe manifest has no lifecycle hooks, and the suspicious agent-configuration behavior is a first-party bootstrap capability that writes bounded read-only project permissions. This is a guarded agent capability, not evidence of malicious install behavior.