bmad_skill
- Python
12
GitHub Stars
2
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill supercent-io/skills-template --skill bmad- SKILL.md4.5 KB
- SKILL.toon1.2 KB
Overview
This skill is an AI-driven agile development framework that orchestrates 26+ specialized agents and 68+ guided workflows to turn informal "vibe coding" into structured, multi-agent development. It follows a four-phase methodology—Analysis, Planning, Solutioning, Implementation—to produce spec-driven results and scale from quick fixes to enterprise systems. The framework adapts planning depth automatically based on project complexity.
How this skill works
The skill coordinates specialized agents (Analyst, Product Manager, Architect, Developer, QA, Barry, and more) to execute phase-appropriate workflows. You initialize the framework at the project level, verify installation via manifest checks, then invoke guided commands for ideation, PRD and tech-spec creation, architecture design, and sprint execution. Party Mode lets you run multiple agents together for complex planning and multi-disciplinary collaboration.
When to use it
- Planning complex software projects that need coordinated multi-agent workflows
- Reducing hallucinations by enforcing spec-driven development before coding
- Scaling development from single features to enterprise-grade systems
- Legacy modernization and incremental brownfield migration
- Cross-team collaboration where humans provide governance and agents perform work
Best practices
- Produce a PRD or Tech Spec before any implementation—SDD first
- Install and run the framework at the project level to keep configuration isolated
- Use Party Mode to combine domain specialists for architecture and planning
- Run QA and quality-gate workflows before merging or deploying
- Humans should review and approve agent outputs; maintain human governance
Example use cases
- Rapid feature delivery: load the Barry agent, generate a lightweight tech spec, then run a sprint plan
- Enterprise project kickoff: run Analysis workflows, create PRD, design architecture, and schedule implementation sprints
- CI/CD automation: perform non-interactive install in a pipeline with specified modules and tools
- Legacy migration: plan incremental modernization with solutioning and implementation tracks coordinated by multiple agents
- Compliance-driven releases: use Planning and QA workflows to ensure requirements and tests meet standards
FAQ
Initialize at project level, verify the manifest, then run phase commands such as /brainstorm-project or /prd for ideation and requirements.
Can I run multiple agents together?
Yes. Use Party Mode (e.g., /party-mode analyst pm architect) or load specific agents with /load-agent [name] to orchestrate collaboration.