How vibepod/vibepod-cli works
VibePod is a CLI for running AI coding agents in isolated Docker or Podman containers. The main workflow is a user invoking vp run, which selects an agent and launches its container; optional configuration, profiles, providers, overlays, ACP, proxy tracking, and Herdr reporting extend that run. Separate commands manage skills and background tasks. The README documents additional analytics and comparison capabilities, but sampled sources do not establish their full implementation or wiring, so those links are intentionally limited.
Main entry points: run.py Run command · docker.py Container runtime · cli.py Command router · task.py Task commands
Components 25
CLI entry
- Command router
src/vibepod/cli.py
Agent execution
- Run command
src/vibepod/commands/run.py - Configuration
src/vibepod/core/config.py - Agent definitions
src/vibepod/core/agents.py - Container runtime
src/vibepod/core/docker.py - Launch preparation
src/vibepod/core/launch.py - Agent container
Run extensions
- Project overlays
src/vibepod/core/overlay.py - Credential profiles
src/vibepod/core/profiles.py - Provider routing
src/vibepod/core/provider_launch.py - ACP adapter
src/vibepod/core/acp.py - HTTP proxy metrics
src/vibepod/core/proxy_filter.py - Session logging
src/vibepod/core/session_logger.py - Herdr reporting
src/vibepod/resources/herdr/herdr-report.js - Local analytics
src/vibepod/core
Management features
- Skills commands
src/vibepod/commands/skills.py - Skills engine
src/vibepod/core/skills_engine.py - Task commands
src/vibepod/commands/task.py - Task registry
src/vibepod/core/tasks.py
Other components
- CLI user
- Docker / Podman
- Model endpoint
- Herdr socket
- Skills container
- ACP editor
Connections 26
- CLI user → Command router: invokes
- Command router → Run command: dispatches
- Command router → Skills commands: dispatches
- Command router → Task commands: dispatches
- Run command → Configuration: reads
- Run command → Agent definitions: selects
- Run command → Launch preparation: prepares
- Run command → Container runtime: launches through
- Container runtime → Docker / Podman: controls
- Docker / Podman → Agent container: runs
- Run command → Project overlays: builds overlay
- Run command → Credential profiles: selects profile
- Run command → Provider routing: prepares routing
- Provider routing → Model endpoint: routes requests
- Run command → ACP adapter: adapts protocol
- ACP editor → ACP adapter: connects
- Run command → HTTP proxy metrics: configures tracking
- HTTP proxy metrics → Local analytics: tracks traffic
- Run command → Session logging: opens session
- Run command → Herdr reporting: enables reporting
- Herdr reporting → Herdr socket: reports state
- Skills commands → Skills engine: invokes
- Skills engine → Container runtime: runs container via
- Container runtime → Skills container: runs
- Task commands → Task registry: records tasks
- Task commands → Container runtime: manages containers