Skills bmad-code-org bmad-method
BMad Method Bundle of 24 skills
- Use it when:
- You want the agent to follow one delivery process, from a vague idea to reviewed, committed code, with as much planning as the change needs.
- It does:
- Gives each phase its own skills and named personas: explore and research, product and UX documents, architecture, then build, review and test, ending with a retrospective per epic.
- Start with:
- The bmad skill's setup, which creates the _bmad/ folder the other skills run their scripts from; ask bmad what to run next.
- You get:
- Markdown files under one output folder, such as briefs, PRDs, UX specs, an architecture spine, plans and research reports, plus local commits.
- For your own skills:
- Smithy (bmad-toolsmith) builds, reviews and slims skills, agents and modules in BMad's format; bmad-eval tests them against the bare model, a rubric or trigger queries.
- Needs:
- uv for BMad's scripts, Node.js for npx, and subagents for builds and reviews.
- Status:
- The default branch is v7, a preview alpha that changes often; the stable v6.12 installs separately, with npx bmad-method install.
TL;DR
BMad Method carries every decision forward as files. The bmad skill's setup creates the _bmad/ folder whose scripts the other skills run, and bmad later answers what to run next. Work is grouped into initiatives, each a ticket tree: the initiative, its epics and their stories. A small change goes straight to bmad-build, which plans, implements, has a subagent review the diff and commits locally; bmad-build-auto does the same unattended. Bigger work gets planning first. bmad-brainstorming, bmad-forge-idea, bmad-deep-recon and bmad-prfaq explore and test ideas; John (bmad-agent-pm) coaches PRDs and other product documents; bmad-ux writes DESIGN.md and EXPERIENCE.md; bmad-architecture records the decisions separately built parts must share. Sofia (bmad-agent-po), in the full install but not listed here, turns those documents into an initiative's requirements, epics and stories. Mary, Sally, Winston and Amelia are personas that stay on and run these skills from a menu. After a build, bmad-code-review, bmad-walkthrough and bmad-qa-generate-e2e-tests check it; bmad-retrospective then judges each finished epic by its own criteria. bmad-party-mode, bmad-advanced-elicitation and bmad-review bring other views to a draft or a change, bmad-project-context keeps AGENTS.md's agent rules verified, and bmad-customize changes a skill through override files. Smithy (bmad-toolsmith) builds your own skills, and bmad-eval measures them.
Summarizes the repo at e734d3a. Its skills have changed by 1,838 words since.
Any agent, with the skills CLI, plus the repo's 10 other skills, among them Sofia (bmad-agent-po), bmad-facilitate and the module records setup needs
npx skills add bmad-code-org/BMAD-METHOD --skill '*' Source on GitHubAll skills from bmad-code-org/bmad-method
- Total Downloads
- 16,600 of bmad, its most installed skill, on skills.sh, Nov 9, 2026
- Repo stars
- 55,140 bmad-code-org/bmad-method
- First seen
- Mar 12, 2026 on skills.sh, its first skill
Its skills, in the order they're used
The ones marked “Only in the bundle” need the rest of it. The others also work alone.
-
Sets up BMad in a project and recommends what to run next, from the help of the BMad modules you have installed.
Sets up and advises on the other BMad skills: its setup and help work from the modules and skills installed beside it, so alone it has nothing to set up, answer from or recommend.
-
Turns the agent into Mary, BMad's business analyst, who stays on until dismissed and runs BMad's research, brainstorming and brief skills.
Its menu runs other BMad skills, mostly bmad-deep-recon and bmad-brainstorming, and it needs the _bmad/ folder that the bmad skill's setup creates.
-
Runs a brainstorming session on any topic, rotating creative techniques to push past 100 ideas, then helps you narrow them down.
Runs its scripts from the _bmad/ folder that the bmad skill's setup creates, and offers that setup when it's missing.
-
Pressure-tests a half-formed idea, one question at a time and with personas probing weak spots, until you act on it or drop it.
Runs its scripts from the _bmad/ folder that the bmad skill's setup creates, and offers that setup when it's missing.
-
Researches the question behind a decision, on the web or in your codebase, and writes a report where every claim cites a source.
Runs its scripts from the _bmad/ folder that the bmad skill's setup creates, and offers that setup when it's missing.
-
Turns the agent into John, BMad's product manager, who coaches PRDs, briefs, business cases and other planning documents out of you.
Hands every document to the repo's bmad-facilitate skill, and needs the _bmad/ folder that the bmad skill's setup creates.
-
Turns the agent into Sally, BMad's UX designer persona, who stays on until dismissed and runs BMad's UX specification workflow.
Its one menu item runs bmad-ux, and it needs the _bmad/ folder that the bmad skill's setup creates.
-
Captures your UX vision in two specs: DESIGN.md for how the product looks and EXPERIENCE.md for how it behaves.
Runs its scripts from the _bmad/ folder that the bmad skill's setup creates, and offers that setup when it's missing.
-
Turns the agent into Winston, BMad's system architect persona, who stays on until dismissed and runs BMad's architecture workflow.
Its one menu item runs bmad-architecture, and it needs the _bmad/ folder that the bmad skill's setup creates.
-
Coaches you through the architecture decisions that keep separately built parts of a system consistent, and records them in a short document.
Runs its scripts from the _bmad/ folder that the bmad skill's setup creates, and offers that setup when it's missing.
-
Turns the agent into Amelia, BMad's test-first senior engineer, who stays on until dismissed and runs BMad's build, test and review skills.
Its menu runs bmad-build, bmad-qa-generate-e2e-tests and bmad-code-review, and it needs the _bmad/ folder that the bmad skill's setup creates.
-
Implements a story, feature or bug fix end to end: plans it, codes it, has subagents review the diff, then commits locally.
Its workflow is printed by a script in the _bmad/ folder that the bmad skill's setup creates, so it does nothing without that setup.
-
Implements one story, feature or bug fix without asking anything: plans, codes, reviews with subagents, commits locally and reports a final status.
Its workflow is printed by a script in the _bmad/ folder that the bmad skill's setup creates, so it does nothing without that setup.
-
Reviews a code change with several independent reviewer subagents, verifies and triages their findings, then applies the fixes you choose.
Its workflow is printed by a script in the _bmad/ folder that the bmad skill's setup creates, so it does nothing without that setup.
-
Walks you through a commit, PR, file or directory one block at a time, in a written narrative, until you call each part done.
Its workflow is printed by a script in the _bmad/ folder that the bmad skill's setup creates, so it does nothing without that setup.
-
Adds end-to-end and API tests for a feature that already works, following your project's existing test setup, and reports bugs they find.
Its workflow is printed by a script in the _bmad/ folder that the bmad skill's setup creates, so it does nothing without that setup.
-
Reviews a finished BMad epic from its tickets, plans, diffs and commits, and writes a retrospective with sourced findings, action items and a verdict.
Its workflow is printed by a script in the _bmad/ folder that the bmad skill's setup creates, and it reviews an epic from BMad's ticket tree, which Sofia (bmad-agent-po, not listed here) splits into tickets and the build skills carry out, so it does nothing without both.
-
bmad-project-context
14.3K total downloadsWrites or refreshes a small, verified block of agent instructions in your repo's AGENTS.md, and records agent mistakes as pitfalls.
-
Reviews a diff, PR or document through several lenses, from adversarial critique to edge cases, test gaps, structure and prose, and reports the findings.
Runs its scripts from the _bmad/ folder that the bmad skill's setup creates, and offers that setup when it's missing.
-
Pushes a draft past its first version: cold-reader subagents critique it with chosen methods, and you accept or reject one merged revision.
Runs its scripts from the _bmad/ folder that the bmad skill's setup creates, and offers that setup when it's missing.
-
Runs a round-table where BMad's agent personas, or personas you define, argue with each other and with you until you call it done.
Runs its scripts from the _bmad/ folder that the bmad skill's setup creates, and offers that setup when it's missing.
-
Writes TOML overrides that change installed BMad agents and workflows, such as a persona, template or extra steps, for your team or yourself.
Only writes overrides for installed BMad skills, into the _bmad/ folder that the bmad skill's setup creates.
-
Turns the agent into Smithy, BMad's toolsmith, who builds skills and agents from a plan you approve, and reviews, trims or packages existing ones.
Needs the _bmad/ folder that the bmad skill's setup creates, whose scripts load its settings and check modules, and offers that setup when it's missing; its evals run through bmad-eval.
-
bmad-eval
7.3K total downloadsBenchmarks a skill against the bare model, another version or a rubric, and tests whether its description fires on the right requests.