CLI Reference activation-first
All commands: 0dai <command> [options]
This page tracks the shipped command registry in cli/npm-package/bin/0dai.js. Run 0dai --help for the live top-level summary, or 0dai <command> --help for per-command flags.
Core Commands
| Command | Description |
|---|
init [--dry-run] | Initialize ai/ layer via API after auth, free activation, and project bind; --dry-run previews files without writing |
sync [--dry-run] [--yes] [--quiet] [--force] | Re-sync configs from server after activation and project bind; full flag list via 0dai sync --help |
project bind [--json] | Bind the current local project to your 0dai account after auth |
project status [--json] | Show local binding status, project id, stack, and origin |
detect | Show detected stack and available/optional CLIs |
reflect [--days N] | Session reflection: delivered tasks, delegation rate per agent, budget spend, model tier breakdown; default 7 days |
models | Model ratings table with tier, score, CLI flag; flags: --fast, --balanced, --deep, --available |
status [--json] | Show maturity, swarm queue, active session |
quickstart | Run auth, init, doctor, and status checks in order |
Doctor & Health
| Command | Description |
|---|
audit | Scan ai/, CLAUDE.md, and agent configs for leaked secrets; exits 1 on critical findings |
doctor [--drift] | Health check + agent auth detection (detects subscription logins, API keys, and CLI presence); explains WHY configs are missing and whether to run init or sync |
validate | Verify required ai/ files are present; prints human-readable summary of missing files; exits 1 on failure |
Auth Commands
| Command | Description |
|---|
auth login [--device] [--no-browser] [--code CODE] [--mcp] | Device code flow — authenticate via browser |
auth mcp [--device] [--no-browser] | Store an MCP token from current auth or device code |
auth logout | Remove credentials |
auth status | Show account, plan, usage, remaining quota, and activation state |
activate free | Claim the free activation license required for API-first init and sync |
activate code <CODE> | Redeem a Pro/Team activation code |
activate status | Show activation id, plan, and current activation status |
redeem <CODE> | Apply a plan upgrade code (format: PLAN-XXXX-XXXX) |
Swarm Commands
Subcommand syntax matches cli/npm-package/lib/commands/swarm.js. Most subcommands gate on a Pro plan.
| Command | Description |
|---|
swarm status | Queue / active / done counts |
swarm add --task "..." --for codex | Create task for agent (--to also accepted) |
swarm delegate --task "..." --to aider | Delegate to another agent (alias of add) |
swarm-run | Add, dispatch, and wait for one swarm task; emits JSON |
swarm budget | Session + daily token spend; model tier breakdown (fast/balanced/deep) |
swarm estimate --goal "..." [--agent X] [--model TIER] [--json] | Server-side cost estimate before dispatch |
swarm quality [--last N] [--details ID] [--json] | Score recent task results via quality_scorer.py |
swarm sessions [--json] | Live session registry table (idle / busy / stale) |
swarm webhook add <url> [--event task_done|task_failed|all] [--secret TOKEN] | Register webhook; URL is positional, event defaults to all |
swarm webhook list | Show registered webhooks |
swarm webhook remove <url> | Remove a webhook (URL positional) |
swarm webhook test [<url>] | Send test payload to a webhook (defaults to first registered) |
Session Commands
| Command | Description |
|---|
session save --goal "..." | Save session for roaming |
session status | Show active session |
session complete | Archive session |
Feedback Commands
| Command | Description |
|---|
feedback push | Send feedback reports to 0dai API |
feedback retry | Retry queued feedback after a failed push |
feedback log --type TYPE --detail TEXT | Log local feedback; types: bug, suggestion, friction, positive |
feedback list | List submitted reports |
Terminal / Multi-session
| Command | Description |
|---|
terminal launch --tool TOOL [-- prompt...] | Launch interactive agent session; tools: codex, claude, gemini, opencode, aider |
terminal list | Show active sessions |
terminal attach <id-prefix> | Re-attach to a running session by id prefix |
terminal kill <id-prefix> | Kill a session by id prefix |
Skills
Skills are slash-commands inside agent CLIs (Claude Code, Codex, Cursor, etc.), not 0dai subcommands. Run them from your agent prompt.
| Skill | Description |
|---|
/build | Run build + test + lint from commands.yaml |
/review | Git diff → tests → decisions → maturity check |
/bugfix | Reproduce → fix → verify → regression test |
/delegate | Save session → create swarm task → handoff |
/feedback | Log bug / suggestion / friction / positive |
/status | Doctor + validate + maturity + swarm |
Global Flags
| Flag | Description |
|---|
--target PATH | Override working directory; applies to all commands |
--version | Show CLI version |
More Commands
Additional shipped surfaces — run 0dai --help for short descriptions and per-command --help for flags: init-existing, run, watch, metrics, heatmap, portfolio, standup, workspace, runner, ci, provider, quota, graph, report, compliance, experience, persona-simulate, receipt, boneyard, gh, loop, tui, paste, play, harvest, delegate, ssh, import claude-code-agents, live, update, security, drift, bundle, replay.