Preview. These maps are our reading of public sources and haven't been checked with the projects' maintainers yet.
Spec-driven development
Tools that make a written spec the center of building with coding agents.
20 projects70% active65% build on a shared standardMapped 2026-10-25
What they agree on
- Is it actively maintained? active (75% of 20)
- How is the main thing licensed? permissive open source (MIT/Apache/BSD) (75% of 20)
- Can it run entirely on your own device or infrastructure, without the maker's servers? yes (80% of 20)
- Do you need an account to use it? no account (75% of 20)
- Who is it for? developers (100% of 20)
- Where do specs live? plain files in the project repository (90% of 20)
- Which coding agents does it work with? many agents (five or more) (79% of 19)
What they argue about
How is it mainly sustained?
Which shared standard or format does it mainly build on?
Once code exists, what is the spec?
After implementation, what built-in step keeps specs and code aligned? (if several, the first that applies)
What format are specs written in?
What form does it mainly ship in?
How is the path from idea to code structured?
Does a human approve the spec or plan before code is written?
By default, once a plan is approved, how is implementation run?
What does it offer teams beyond sharing files in git? (if both, pick the paid product)
Projects
| Project | Status | What it is |
|---|---|---|
| *codeplain | Active | Spec-as-source platform: agents help write specs in the ***plain language and the hosted renderer generates tested code; specs, not code, a… |
| Agent OS | Active | Brian Casel's system for discovering and injecting codebase standards and shaping specs; v3 defers spec writing to agents' plan mode and dr… |
| AGENTS.md | Active | Open file convention: a README for agents with build, test and style instructions, read by most coding agents. Standing context, not a feat… |
| Backlog.md | Active | Markdown-native task manager and Kanban for git repos, structured around three human checkpoints (spec, plan, code) with acceptance criteri… |
| BMAD Method | Active | Free 'Agile AI-Driven Development' method: role-based agents and workflows (analyst, PM, architect, dev, QA) that carry briefs, PRDs, archi… |
| cc-sdd | Active | Kiro-style requirements/design/tasks workflow packaged as 17 Agent Skills for 8 agents, with discovery routing, EARS requirements and long-… |
| Claude Code Spec Workflow | Quiet | Earlier Claude Code-only slash-command workflow (Requirements -> Design -> Tasks -> Implementation, plus a bug-fix flow), superseded by the… |
| Conductor (Google) | Maintenance | Google's context-driven development plugin: product/tech-stack context files, then per-track spec.md and plan.md, implementation with check… |
| GitHub Spec Kit | Active | GitHub's open-source toolkit (specify CLI plus slash-command skills) for constitution -> specify -> plan -> tasks -> implement -> converge,… |
| GSD (Get Shit Done) / GSD Core | Active | Meta-prompting and context-engineering framework running discuss -> plan -> execute -> verify -> ship per phase, with heavy work in fresh-c… |
| Intent (Augment) | Active | Free macOS desktop workspace for multi-agent work: a Coordinator writes a living spec, Implementor agents run tasks in parallel, a Verifier… |
| Kiro | Active | AWS's spec-driven agentic IDE (VS Code-based), now one agent harness across IDE, CLI, web (PR-opening background agent), mobile and 'Crew';… |
| OpenSpec | Active | Lightweight, brownfield-first spec layer: each change gets proposal/specs/design/tasks; archiving merges spec deltas into living specs. Wor… |
| PRP (Product Requirement Prompts) | Active | Rasmus Widing's methodology: a PRP = PRD + curated codebase context + validation commands; skills for PRD, plan, implement, review and an a… |
| Spec Kitty | Active | Spec -> plan -> tasks -> review -> accept -> merge CLI for 'governed software factories': work packages in lanes, parallel agents in git wo… |
| Spec Workflow MCP | Maintenance | MCP server for Requirements -> Design -> Tasks specs with a local web dashboard, VS Code extension, approval workflow and implementation lo… |
| Superpowers | Active | Jesse Vincent's skills framework and development methodology (brainstorm, plan, subagent-driven development, TDD) installed as a plugin in… |
| Taskmaster AI | Maintenance | PRD-to-task-graph manager (CLI + MCP) that parses a requirements doc into dependent tasks for Cursor, Claude Code and others; now documente… |
| Tessl | Changed direction | Started as the 'spec-as-source' company (Tessl Framework + Spec Registry); now a platform for writing, evaluating and distributing agent sk… |
| Traycer | Changed direction | Desktop app that orchestrates many coding agents in parallel worktrees with shared context, specs/tickets as persistent artifacts, agent-to… |
Where they meet
Not mapped yet.
Follow
RSS for this community: projects added, status changes, and later polls and discussions.
Follow the whole community: OPML of every project's own release feed (where we found one) · OPML of every project's Poll Position feed.
The poll
Every project above was answered against these questions. Core questions are shared by all communities; the rest are this community's own.
- Is it actively maintained? core
active · active, but changed direction (pivot, fork, or rename) · maintenance or slowing · dormant, sunsetting, or archived - How is the main thing licensed? core
permissive open source (MIT/Apache/BSD) · copyleft or source-available with restrictions · open parts around a closed core · proprietary or closed source - How is it mainly sustained? core
individual side project, no revenue · individual with revenue (paid app, tier, ads, or donations) · community, foundation, or research · company with a paid product, tier, or ads · company or startup with a free product (no revenue yet) · big company side project - Can it run entirely on your own device or infrastructure, without the maker's servers? core
yes · yes, but it calls a third-party service you choose (your own API key or account) · partly (some features need the maker's servers) · no, it needs the maker's servers · not applicable - Do you need an account to use it? core
no account · optional account · account required - Who is it for? core
developers · non-developers · both - Is there public evidence it's built with AI coding agents? core
yes, the repo shows it (agent instruction files or agent co-authored commits) · yes, the maker says so · no public evidence - Which shared standard or format does it mainly build on? core
Agent Skills (SKILL.md) · Model Context Protocol (MCP) · AGENTS.md · the Kiro spec layout (requirements in EARS, design, tasks files) · none of these (its own format or interface) · not applicable - Once code exists, what is the spec? value
the source: changes start in the spec and code follows · a living document kept current alongside the code · a planning scaffold: code becomes the truth after implementation · there is no feature spec; it supplies standing context or rules - After implementation, what built-in step keeps specs and code aligned? (if several, the first that applies) approach
code is regenerated from the spec · a step folds changes back into the specs · a step checks code against the spec (tests, review, or verifier) · nothing built in - What format are specs written in? approach
markdown from templates or free-form · markdown with a required requirement syntax (e.g. EARS, SHALL/WHEN scenarios) · a structured schema or dedicated spec language - Where do specs live? value
plain files in the project repository · the tool's own app, database, or cloud · both, synced - Which coding agents does it work with? approach
only its own built-in agent · one to four named third-party agents · many agents (five or more) - What form does it mainly ship in? approach
prompts, skills, or slash commands installed into an agent (even if a CLI installs them) · a CLI or MCP server that itself reads and writes the spec or task files · a standalone app, IDE, or hosted platform · a file convention only - How is the path from idea to code structured? approach
no fixed phases · a fixed sequence of phases (e.g. requirements, design, tasks) · adaptive: the process scales to the size of the change - Does a human approve the spec or plan before code is written? value
yes, the workflow waits for approval · review is recommended but not enforced · no approval step - By default, once a plan is approved, how is implementation run? value
the human drives it, phase by phase or task by task · agents run the whole plan with automated review or verification loops · it does not run implementation - What does it offer teams beyond sharing files in git? (if both, pick the paid product) approach
nothing beyond git · team features in the open-source tool (shared spec stores, multi-developer or multi-agent coordination) · team features in a paid hosted product
Sources
Picked by hand (Phase 0); see the case study entry 'Whose communities?'.
77 sources across 20 projects, read on 2026-10-25. Each project page lists its own. Poll version: core-v4.