prodigy 0.4.4

Turn ad-hoc Claude sessions into reproducible development pipelines with parallel AI agents
Documentation
# implement.yml - Configuration for implementing specifications with validation

# Commands to execute for each specification
commands:
  # Step 1: Implement the specification
  - claude: "/prodigy-implement-spec $ARG"
    commit_required: true
    validate:
      claude: "/prodigy-validate-spec $ARG --output .prodigy/validation-result.json"
      result_file: ".prodigy/validation-result.json"  # Where Prodigy reads validation results from
      threshold: 100  # Spec must be 100% implemented (default)
      on_incomplete:
        claude: "/prodigy-complete-spec $ARG --gaps ${validation.gaps}"
        max_attempts: 5
        fail_workflow: false  # Continue even if we can't reach 100%
        commit_required: true  # Require commit to verify fixes were made

  # Step 2: Run tests to verify implementation
  - shell: "just test"
    on_failure:
      claude: "/prodigy-debug-test-failure --spec $ARG --output ${shell.output}"
      max_attempts: 5
      fail_workflow: false  # Continue workflow even if tests can't be fixed

  # Step 3: Run linting and formatting after implementation
  - shell: "just fmt-check && just lint"
    on_failure:
      claude: "/prodigy-lint ${shell.output}"
      max_attempts: 5
      fail_workflow: false

# Custom merge workflow - ensures clean integration back to master
merge:
  # Step 1: Merge master into worktree
  - claude: "/prodigy-merge-master"  # Handles merge conflicts if any

  # Step 2: Run CI checks and fix any issues
  - claude: "/prodigy-ci"

  # Step 3: Merge worktree back to original branch
  - claude: "/prodigy-merge-worktree ${merge.source_branch} ${merge.target_branch}"