Deterministic local execution runtime for agentic software workflows with native BMAD agile methodology. 12+ specialized agents, 34+ workflows, Party Mode included.
All critical gaps addressed. Ready for enterprise deployment with comprehensive testing and documentation.
Complete SDLC automation with Git integration, vector store, skill governance, and multi-agent coordination.
AI agents write, refine, and optimize code based on natural language prompts. Continuous improvement through feedback loops.
QA agents automatically identify bugs and Dev agents fix them without human intervention.
Automated PR/MR creation, CI integration, and auto-merge with conflict resolution.
PostgreSQL + pgvector backend for scalable knowledge management and similarity search.
Built-in observability with core types, configuration, and APM integration guides for production monitoring.
Ed25519 signatures, trusted registry, and audit logging for supply chain security.
Parallel execution with conflict resolution and state management for complex workflows.
Native BMAD-METHOD integration with 12 specialized agents (PM, Architect, Developer, UX, QA), 34+ workflows, and Party Mode for multi-agent collaboration.
From CLI to execution engine, platform tiers, and office simulation — a clear, layered architecture built for deterministic automation.
The runtime separates orchestration, execution, and governance. Each tier can evolve independently while staying fully observable.
Repeatable runs with state stores, replay, and strict resource budgets.
Formal verification, adversarial testing, and cryptographic commitments.
Telemetry, cost tracking, tenant isolation, and marketplace distribution.
Direct browser control for AI agents. Web scraping, automated testing, and end-to-end workflows.
800 tokens/page with text extraction, 5-13x cheaper than screenshots
Run without a window or with visible Chrome for debugging
Run multiple parallel Chrome processes with isolated profiles
Advanced stealth injection and IDPI protection enabled by default
# Launch browser and navigate
Skill: browser_automation
action: navigate
url: "https://github.com/pinchtab/pinchtab"
# Get interactive elements
action: snapshot
# Click element (by ref, not coordinates)
action: click
element_id: "e5"
# Fill form field
action: fill
element_id: "e3"
value: "user@example.com"Native BMAD methodology with 10+ pre-built workflows for the complete software development lifecycle.
Plan your sprint with backlog refinement, capacity planning, and sprint commitment.
# Run Sprint Planning
cargo run -- --workflow sprint-planning \
--task "Plan sprint 12 with 40 story points"
# With deterministic replay
cargo run -- --workflow sprint-planning \
--task "Plan sprint 12" \
--save-replay sprint12.jsonFull workflow for developing new features with requirements, design, implementation, and testing.
# Run Feature Development
cargo run -- --workflow feature \
--task "Build user authentication with JWT"
# Scale adaptive (auto-adjusts)
cargo run -- --workflow feature \
--task "Build payment system" \
--scale enterpriseDaily sync meeting for sprint progress with yesterday/today/blockers format.
# Run Daily Standup
cargo run -- --workflow standup \
--task "Daily standup for sprint 12"
# Generate standup notes
cargo run -- --workflow standup \
--task "Generate standup summary" \
--output notes.mdWorkflow for releasing versions with testing, security review, deployment, and monitoring.
# Run Release Workflow
cargo run -- --workflow release \
--task "Release version 2.0.0"
# Dry run (no actual deployment)
cargo run -- --workflow release \
--task "Release v2.0.0" \
--dry-runMultiple AI agents collaborate in a single session. Perfect for complex tasks requiring diverse expertise.
# Start Party Mode - Feature Team
cargo run -- --party-mode feature-team \
--task "Build API for user management"
# Party Mode - Full Stack Team
cargo run -- --party-mode fullstack \
--task "Build user dashboard"
# Party Mode - Platform Team
cargo run -- --party-mode platform \
--task "Design microservices architecture"From specification to production in four simple steps.
Define your requirements in a simple markdown specification file.
Specialized AI agents write, test, and refine code automatically.
Self-healing QA agents identify and fix bugs autonomously.
Your application is ready with full telemetry and trace timeline.
Get up and running in minutes with these simple commands.
# Clone the repository git clone https://github.com/truongnat/agentic-sdlc.git cd agentic-sdlc # Bootstrap the environment ./scripts/bootstrap.sh # Run workflow doctor cargo run -- workflow doctor # Run a workflow cargo run -- --workflow valid_flow.md # Or install globally cargo install --path . antigrav workflow doctor
# Run with BMAD feature workflow cargo run -- --workflow feature --task "Build user auth" # Run sprint planning workflow cargo run -- --workflow sprint-planning --task "Plan sprint 12" # Use Party Mode for multi-agent cargo run -- --party-mode feature-team
Pre-built skill packages for specialized development domains.
Build and deploy AI models with integrated ML workflows.
Cloud-native development and infrastructure automation.
Security scanning, penetration testing, and compliance.
Data pipeline management and ML model evaluation.
Healthcare software development and compliance tools.
Environmental monitoring and sustainability solutions.