safe-agentic-workflow

🥈Silver

Production-validated SAFe multi-agent development methodology with Claude Code. Treat AI agents as specialized team members (11 roles: BSA, Architect, QAS, etc). Complete whitepaper + working template.

5380Updated 1mo ago

Intermediate30min to implementautomation

Saves ~240 min per use

Quick InstallView Source

git clone https://github.com/bybren-llc/safe-agentic-workflow.git

Works with:

Claude

Overview

About This Skill

How to Use

1. **Initialize the Team:** Start by defining your [PROJECT_NAME], [SPRINT_NUMBER], and [BACKLOG_ITEMS] in the prompt template. Use your project management tool (e.g., Jira, Azure DevOps) to extract the current sprint backlog. 2. **Customize Roles:** Ensure all 11 SAFe roles are represented in your team. If a role is missing (e.g., no Data Engineer), assign those tasks to the closest match (e.g., QAS or Developer) and document the gap. 3. **Run the Sprint:** Use the generated sprint plan in your daily standups (e.g., in Slack/Teams with a bot posting the standup script). Update the plan in real-time if blockers arise (e.g., "Payment provider API changed—DevOps to investigate impact"). 4. **Validate Outputs:** After each task, have the assigned role verify the output against the DoD checklist. Use tools like SonarQube for code quality, LoadRunner for performance, and Figma for UX reviews. 5. **Retrospective:** At the end of the sprint, use the plan to identify what worked (e.g., "UX prototype delivered early") and what didn’t (e.g., "Security audit delayed by 2 days"). Adjust the template for the next sprint.

Setup & Installation

Quick Install

No install command available. Check the GitHub repository for manual installation instructions.

Alternative Install (Git Clone)

git clone https://github.com/bybren-llc/safe-agentic-workflow

Requirements

Claude Code or compatible AI agent
Works with: Claude

Quick Start Guide

Install the Skill

Copy the install command above and run it in your terminal.

Open Your AI Agent

Launch Claude Code, Cursor, or your preferred AI coding agent.

Try It Out

Use the prompt template or examples below to test the skill.

Customize

Adapt the skill to your specific use case and workflow.

Usage Examples

Prompt Template

Act as a SAFe multi-agent development team using the 11 predefined roles: BSA (Business Solutions Analyst), Architect, QAS (Quality Assurance Specialist), Developer, DevOps Engineer, Product Owner, Scrum Master, Release Train Engineer, System Team Member, UX Designer, and Data Engineer. Follow the SAFe Agile Release Train (ART) methodology for the project: [PROJECT_NAME]. The current sprint is [SPRINT_NUMBER] (start date: [SPRINT_START_DATE], end date: [SPRINT_END_DATE]). Current backlog items: [BACKLOG_ITEMS]. Your task is to: 1) Decompose [BACKLOG_ITEMS] into SAFe user stories and enabler tasks, 2) Assign roles to each task based on expertise, 3) Define dependencies and risks, 4) Generate a sprint plan with daily standup scripts, and 5) Create a definition of done (DoD) checklist. Use the SAFe PI Planning template structure. Respond with a structured report in Markdown format.

Example Output

```markdown
# SAFe Multi-Agent Sprint Plan: Q3 2024 E-Commerce Platform - Sprint 2
**Project:** Next-Gen Checkout Experience | **Sprint Dates:** July 8 - July 19, 2024
**Team:** ART Alpha (11 members) | **PI:** Q3 2024

## 1. Backlog Decomposition & Role Assignment

### User Story 1: Implement One-Click Checkout
- **BSA (Business Solutions Analyst):** Validate business value with stakeholders. *Output:* Acceptance criteria: "Users can complete purchase in ≤3 clicks with 95% success rate."
- **UX Designer:** Design micro-interaction flows. *Output:* Figma prototype of new checkout flow with accessibility compliance.
- **Developer:** Implement frontend components. *Output:* React components for OneClickCheckout.tsx with Jest unit tests.
- **QAS (Quality Assurance Specialist):** Test edge cases. *Output:* Test plan covering: slow network conditions, payment failures, and guest checkout scenarios.
- **DevOps Engineer:** Containerize and deploy. *Output:* Dockerfile and Kubernetes HPA configuration for auto-scaling.

### Enabler Task: Security Audit of Payment Gateway
- **Architect:** Review current security posture. *Output:* Threat model diagram and gap analysis report.
- **Data Engineer:** Analyze transaction logs. *Output:* SQL queries to identify anomalous payment patterns.
- **System Team Member:** Set up staging environment. *Output:* Terraform scripts to replicate production infrastructure.

## 2. Dependencies & Risks
- **Critical Path:** UX Designer must deliver prototype by Day 3 (July 10) to avoid blocking Developers.
- **Risk:** Third-party payment provider (Stripe) API changes may impact integration timeline. *Mitigation:* QAS to validate API compatibility in sandbox environment.

## 3. Sprint Plan (10 business days)

| Day | Focus Area | Standup Script | Owner |
|-----|------------|----------------|-------|
| 1 (Jul 8) | Sprint Planning & Backlog Refinement | "Yesterday: BSA finalized acceptance criteria. Today: UX to deliver prototype. Blockers: None." | Scrum Master |
| 2 (Jul 9) | Prototype Review & Security Audit Kickoff | "Yesterday: UX delivered Figma prototype. Today: Architect to review threat model. Blockers: Waiting on Data Engineer for logs." | Release Train Engineer |
| ... | ... | ... | ... |
| 10 (Jul 19) | Sprint Review & Retrospective | "Yesterday: All user stories met DoD. Today: Prepare demo for stakeholders. Blockers: None." | Product Owner |

## 4. Definition of Done (DoD) Checklist
- [ ] Code reviewed by at least 2 team members (Architect + Developer)
- [ ] Unit tests pass (80% coverage minimum)
- [ ] Security scan passes (SonarQube: no critical vulnerabilities)
- [ ] Performance tested (LoadRunner: <2s response time under 1000 concurrent users)
- [ ] Documentation updated (Confluence: README.md + API docs)
- [ ] Stakeholder sign-off (PO + BSA)

**Total Story Points:** 45 (13 stories) | **Team Velocity:** 42 (historical average: 40-45)
```