What problem does it solve? Operating the mj-agent independent Docker Compose stack (mj-agent, mj-agent-postgres, mj-agent-redis) requires choosing the correct -f file chain per profile, running pre-checks against the mj-system network, and diagnosing failures like port conflicts, unhealthy containers, or Harbor pull errors. This Skill standardizes the full up/ps/logs/down lifecycle so operators avoid mis-loaded profiles and destructive mistakes. ## Core Features & Use Cases - Profile-aware lifecycle commands: Runs up, ps, logs, and down with explicit -f chains for dev (base + override), test, and prod overlays per ADR-025 4-file layering. - Pre-flight validation: Verifies mj-system-backend-network exists, mj-system biz postgres is healthy, .env secrets are present, and host ports 8001/5433/6379 are free before starting. - Troubleshooting matrix: Maps common failures (port conflicts, healthcheck timeouts, network not found, Harbor pull denied, proxy-related curl 502) to concrete fixes. - Use Case: A developer says "起 mj-agent compose" on a test host; the Skill runs pre-checks, executes the test-profile up command, verifies healthchecks and port listening, and reports a structured lifecycle report. ## Quick Start Ask the agent to start the mj-agent compose stack in dev profile and verify all three containers are healthy.
