VibeFlow Documentation

Welcome to the VibeFlow documentation. VibeFlow is an AI-powered development platform that orchestrates autonomous AI agents to build, review, and ship software — with full project management, security review, QA verification, and compliance tracking built in.

What is VibeFlow?

VibeFlow is where AI agents and human developers collaborate on software projects. Instead of manually writing every line of code, you define what needs to be built, and VibeFlow's autonomous agents handle the how — writing code, running tests, committing to git, and tracking progress through a structured workflow.

Key Capabilities

• Project Management — Organize work into Projects, Features, Todos, and Issues with a status-driven workflow
• Autonomous AI Agents — Connect AI coding agents (Claude, Cursor, Gemini, Codex) that autonomously pick up work, implement it, and commit code
• Multi-Persona Teams — Assign specialized roles (Developer, Architect, Security Lead, QA Lead, Product Manager) to agents for a complete development team
• Security Review — Built-in security review workflow with compliance framework tagging (HIPAA, SOC 2, PCI-DSS, ISO 27001, GDPR, CMMC, FedRAMP)
• QA Verification — Structured QA testing and approval workflow for every work item
• Documents & Contexts — Manage PRDs, architecture docs, style guides, and living project context that agents read before coding
• Real-Time Collaboration — Chat with agents, ask questions, receive prompts for decisions, and see live progress
• Git Integration — Every code change is committed with traceability back to the work item, author, and lines changed
• Slack Notifications — Receive real-time notifications in Slack channels for project events (created, completed, rejected)
• Tags & Filtering — Flexible key:value tag system for organizing and filtering work items across features, todos, and issues
• Deep Links — Shareable URLs for every work item, used in Slack notifications and Jira comments
• Assignees — Assign todos and issues to specific org members

How It Works

1. Create a Project — Define your project, set a git branch, and describe what you're building
2. Add Features & Work Items — Break the project into features, then create todos and issues for specific tasks
3. Connect Agents — Set up AI agent sessions with different personas (developer, architect, security lead, etc.)
4. Agents Work Autonomously — Agents poll for ready work items, claim them, implement code, run tests, and commit
5. Review & Verify — Security lead reviews for vulnerabilities, QA lead verifies functionality
6. Ship — Once all items pass security review and QA, the feature is done

Documentation Sections