CodeCrew compared with Spec Kit, BMAD, OpenSpec and GSD¶
If you have been reading about Spec Kit, BMAD, OpenSpec and GSD, you have met tools for planning and building with an AI coding agent. CodeCrew works on a neighbouring problem: which agent does the work, which one checks it, and where the record lives. Here is what each is for, and when to pick it.
At a glance¶
| Tool | Good for | Where it writes things down |
|---|---|---|
| Spec Kit | Turning a written spec into a plan, tasks and code | Files such as spec.md, plan.md and tasks.md |
| BMAD Method | Planning sized to the work, from several angles | Briefs, specs and architecture docs |
| OpenSpec | A proposal, specs and tasks for each change | An openspec/ folder of specs and changes |
| GSD | A phase loop run by fresh-context subagents | Files such as STATE.md and CONTEXT.md |
| CodeCrew | One agent builds, a separate one reviews | GitHub issues and pull requests |
Spec Kit¶
GitHub's open-source toolkit “gives AI coding agents structured processes, reusable templates, and documented outcomes”. Its core process is spec-driven development: you write what to build and why, and your agent turns it into a plan, tasks and code, one reviewed step at a time. Pick it when a written spec should drive the build. (repository, quickstart)
BMAD Method¶
BMad Method sets out to “turn an idea or change request into working software without giving up the thinking”. It sizes the process to the change and brings in product, architecture, UX, development and testing perspectives when they help. Pick it when you want help deciding what to build, not only building it. (repository, docs)
OpenSpec¶
OpenSpec is “Spec-driven development (SDD) for AI coding assistants”. Each change gets a proposal, specs, a design and a task list. It describes itself as “built for brownfield not just greenfield”. Pick it when you want a spec for each change, in new or existing code. (repository, docs)
GSD¶
GSD is where CodeCrew began. It runs each phase through a loop of discuss, plan, execute, verify and ship, with the heavy work in fresh-context subagents to avoid what it calls context rot. CodeCrew started as an essay written from inside a GSD project, which found GSD's value in “externalized state, independent verification, human gates on judgment calls, atomic commits”, and puts that core on GitHub. The original GSD repository is archived, and says the project now continues as GSD Core. Pick GSD when you want a complete phase loop inside your coding agent.
CodeCrew¶
CodeCrew does not write specs for you. It gives each agent its own GitHub identity, lets a different agent review the work, and refuses to merge until the checks and the review are in. Pick it when you want separate agents to build and to check, on GitHub.
Where they overlap¶
Planning first, mostly. Spec Kit, OpenSpec and GSD plan before code. BMAD sizes it: clear changes can go straight to build. CodeCrew's task start refuses a task with no plan, however the agent arrived at it.
More than one coding tool. Each works with several coding tools: Spec Kit (integrations), BMAD (install), OpenSpec (supported tools), GSD Core (installer), and CodeCrew, whose role instructions are plain files any harness can load. Supporting several tools is not the same as having a second agent check the first: in CodeCrew the review comes from a separate session, which can hold a GitHub identity of its own.
Next step¶
Try CodeCrew on a project of your own: Your first milestone.
Tried it? Tell us how it went
What did you try, how far did you get, and what stopped you? Tell us in “I tried it”, on GitHub Discussions.