Usage

Helm is invoked inside Claude Code via the /helm command. The CLI manages initialization and state display.

Commands

# Initialize a new project
helm init
helm init --name my-api --type brownfield --lang pt-BR
helm init --force                    # reinitialize existing project

# Check pipeline status
helm status                          # colored dashboard
helm status --short                  # one-line summary
helm status --json                   # machine-readable output

# Resume from where you left off
helm resume                          # show context + next steps
helm resume --json                   # structured output

# Checkpoint session state
helm save                            # validate + checkpoint
helm save --message "before refactor"
helm save --force                    # skip recent checkpoint warning

# Update to latest version
helm update

# Other
helm version
helm help

Global Flags

FlagDescription
--no-colorDisable colored output

Color is automatically disabled when piping output or when NO_COLOR environment variable is set.

Claude Code commands

Inside Claude Code, these commands control the orchestrator:

/helm              # Start or resume the pipeline
/helm status       # Show project dashboard
/helm resume       # Reload session and continue
/helm exit         # Save state and stop
/helm stop         # Same as exit
/helm help         # Show available commands

Daily workflow

Starting a new project

  1. Run helm init in your project directory
  2. Open Claude Code and type /helm
  3. Describe your project. Helm routes to the first agent
  4. Follow the pipeline through each phase

Resuming work

  1. Open Claude Code and type /helm
  2. Helm reads your session and shows where you left off
  3. Choose to continue, review, or check status

Quick flow

For small, well-defined tasks (bug fixes, hotfixes):

  1. Type /helm and describe the fix
  2. Helm detects it’s a quick task and offers a fast track
  3. Fast track: Research (quick) → Build → Verify → Ship
  4. Skips: Planning, Architect, Roadmap, Breakdown, Review

Execution profiles

ProfileBehavior
ExploreRead-only. Agents analyze and suggest but perform no writes.
GuidedDefault. Agents propose actions, user confirms before writes.
AutonomousAgents execute without confirmation when gate scores >= 95%.

If quality drops below 95% during autonomous execution, Helm downgrades to guided mode.

Session files

FilePurposeCommitted
helm.yamlProject config (version, type, language)Yes
.helm/session.yamlRuntime state (current agent, scores, timestamps)No
.helm/handoffs/*.mdAgent handoff documentsNo
.helm/artifacts/Agent output artifactsNo