A Rovo-powered supreme intelligence that observes Jira and Confluence, plans execution like a race engineer, and safely turns intent into action with full auditability.

Pit Crew SI is a Forge app that acts as your team's intelligent assistant, continuously monitoring Jira issues and Confluence spaces to:

• Triage work automatically - Analyzes issues and suggests priorities, labels, and actions
• Propose sprint plans - Generates data-driven sprint plans based on priorities and capacity
• Draft execution briefs - Creates detailed task breakdowns for complex work items
• Execute safe actions - Optionally performs approved actions like labeling, prioritization, and subtask creation
• Maintain full auditability - Logs every action with explanation and revert steps

• Automatic issue analysis using content-based heuristics
• Priority determination based on keywords and context
• Suggested labels based on technical areas (api, ui, database, performance, etc.)
• Effort estimation and actionability assessment

• Action validation - Ensures actions are in the allowed list
• Impact assessment - Evaluates the scope of each action
• Reserved label protection - Prevents unauthorized use of critical labels
• High-priority safeguards - Requires approval for critical/blocker priorities
• Assignment restrictions - Requires approval before assigning work to team members

Dry-Run Mode (Default)

• Analyzes and proposes actions without executing
• Safe for testing and validation
• All actions logged as "proposed"

Execute Mode

• Performs approved actions on Jira issues
• Respects guardrails and approval settings
• Full audit trail maintained

All actions include:

• Clear explanation of what was done
• Step-by-step revert instructions
• Original state tracking (where applicable)
• One-click revert capability

• Complete history of all proposed and executed actions
• Status tracking (proposed, approved, executed, failed, reverted)
• Timestamp and executor information
• Explainable AI - every action includes reasoning

• Analyzes backlog issues
• Prioritizes based on multiple factors
• Respects capacity constraints
• Identifies risks and dependencies

• Breaks down complex tasks into steps
• Provides technical context
• Links related Jira issues and Confluence pages
• Estimates time for each subtask

pitcrew-si/
├── src/
│   ├── index.ts              # Main resolver and API endpoints
│   ├── types.ts              # TypeScript type definitions
│   ├── storage.ts            # Storage service for persistence
│   ├── observers/
│   │   ├── jira-observer.ts      # Monitors Jira issue events
│   │   └── confluence-observer.ts # Monitors Confluence page events
│   ├── engine/
│   │   └── triage.ts         # Triage and planning logic
│   ├── actions/
│   │   └── executor.ts       # Action execution engine
│   └── utils/
│       └── guardrails.ts     # Safety checks and validations
├── static/
│   └── dashboard/
│       └── index.html        # Dashboard UI
├── manifest.yml              # Forge app manifest
├── package.json              # Dependencies
└── tsconfig.json             # TypeScript configuration

• Node.js 20.x or later
• Forge CLI installed: npm install -g @forge/cli
• Atlassian account with admin access to a Jira/Confluence site

1. Clone the repository

   git clone https://github.com/wildhash/pitcrew-si.git
   cd pitcrew-si

2. Install dependencies

   npm install

3. Build the app

   npm run build

4. Login to Forge

   forge login

5. Deploy the app

   forge deploy

6. Install the app on your site

   forge install

   Select your Jira product and follow the prompts.

1. Navigate to your Jira instance
2. Click on "Apps" in the top navigation
3. Select "Pit Crew SI Dashboard"
4. The dashboard will display action logs and controls

When a new issue is created or updated in Jira, Pit Crew SI automatically:

1. Analyzes the issue content
2. Determines appropriate priority
3. Suggests relevant labels
4. Proposes actions (if applicable)
5. Logs the triage result

Use the resolver API to:

// Get triage result for an issue
const result = await invoke('getTriageResult', { issueKey: 'PROJ-123' });

// Generate sprint plan
const plan = await invoke('generateSprintPlan', {
  issues: [...],
  goals: ['Complete API migration', 'Fix critical bugs'],
  capacity: 40
});

// Generate execution brief
const brief = await invoke('generateExecutionBrief', {
  issue: {...},
  relatedPages: [...]
});

// Execute an action
const log = await invoke('executeAction', { proposal: {...} });

Update the configuration through the dashboard or API:

await invoke('updateConfig', {
  config: {
    mode: 'execute',           // 'dry-run' or 'execute'
    autoApprove: false,        // Require manual approval
    guardrailsEnabled: true,   // Enable safety checks
    auditLogEnabled: true,     // Log all actions
    maxActionsPerRun: 10,      // Limit actions per execution
    allowedActions: [          // Permitted action types
      'label-issue',
      'set-priority',
      'create-subtask'
    ]
  }
});

1. Create a new bug: "API endpoint returning 500 errors for user authentication"
2. Pit Crew SI automatically:
   • Sets priority to "high"
   • Adds labels: "bug", "api", "security"
   • Proposes creating subtasks for investigation and fix
3. Review proposals in dashboard
4. Approve and execute (if in execute mode)

1. Have 20+ issues in your backlog
2. Invoke generateSprintPlan with goals and capacity
3. Review the prioritized list
4. Pit Crew SI identifies:
   • Issues that fit in capacity
   • Potential risks
   • Dependencies between issues

1. Select a complex feature issue
2. Generate execution brief
3. Review the breakdown:
   • Clear objectives
   • Step-by-step tasks with estimates
   • Technical context
   • Related documentation

npm run build

npm run lint

npm run format

forge tunnel

This starts a local tunnel for development with hot reload.

• All actions validated before execution
• Reserved labels require manual approval
• High-impact actions require approval
• Rate limiting on actions per run

• Every action logged with full context
• Timestamps and executor tracking
• Explanation for AI decisions
• Revert instructions provided

All supported actions can be reverted:

• Label additions can be removed
• Priority changes can be reversed
• Created subtasks can be deleted
• Assignments can be changed back

We welcome contributions! Please:

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests if applicable
5. Submit a pull request

MIT License - see LICENSE file for details

Built with:

• Atlassian Forge
• Rovo Agents
• TypeScript
• Love for developer productivity ❤️

For issues, questions, or suggestions:

• Open an issue on GitHub
• Contact the Pit Crew SI team
• Check the Forge documentation

Built with ❤️ for developers by developers. May your sprints be smooth and your bugs be few! 🏎️💨