Skip to main content

README

SpecWeave

AI-assisted development, under control.
One folder per unit of work. Six commands. An append-only ledger any tool can write to.

npm10 skills10+ production appsMITDiscord


Install

npm install -g specweave        # Node.js 20.12.0+
cd your-project
specweave init .

The loop

#CommandClaude Code skillWhat it does
1specweave create-increment "<title>"/sw:incrementspec.md with Problem, Scope, numbered ACs and an Approach. Approved before any code.
2specweave task nextclaimdone --run "<test>"/sw:doWork the tasks. done refuses a failing test command.
3specweave verifyRuns your test/lint/build; writes reports/verify.json.
4/sw:reviewFresh-context adversarial review; every finding cites path:line.
5specweave complete <id>/sw:doneCloses. Blocks only on a green verify.json (or --reason).
6specweave handoff/sw:handoffA portable one-page doc so any other tool can pick the work up.

Everything lives in .specweave/increments/NNNN-slug/: spec.md, tasks.md (+ the rendered SW:BOARD), ledger.jsonl, handoff.md, reports/.

The ten skills

The CLI is the product and it runs in any AI tool or in CI. These ten skills are the Claude Code wrappers over it — nothing else ships in the plugin, so nothing goes stale.

SkillUse it for
sw:brainstormFramed alternatives, ending in a pick.
sw:incrementPlan the work: spec.md with ACs, plus tasks.md.
sw:doClaim a task, implement it, close it with evidence.
sw:autoThe same loop, unattended, until the tasks run out.
sw:teamA worktree per agent, claims arbitrated by the ledger.
sw:reviewFresh-context adversarial pass; findings cite path:line.
sw:qaRisk-scored assessment, blockers, verdict.
sw:doneVerify, review check, specweave complete.
sw:syncGitHub, Jira and Azure DevOps: push, pull, status, setup.
sw:handoffA one-page, secret-scrubbed continuation doc.

Five longer procedures (tdd-cycle, e2e, debug, diagrams, release-expert) live in skills-optional/ and install on demand with vskill.

Upgrade to 2.0

npm i -g specweave@2
specweave update

specweave update is idempotent: it rewrites the managed sections of CLAUDE.md/AGENTS.md while preserving your own, migrates config.json in one pass, writes the .gitignore/.gitattributes entries 2.0 needs, and keeps backups under .specweave/backups/. Then run specweave doctor.

2.0 removed a lot on purpose — the three-report closure pipeline, auto-generated living docs, 34 never-invoked skills, the queued sync mode. See What was removed, and why.

There is no alias routing: a 1.x slug simply has no skill behind it. The old-to-new map — every removed skill, the skill or CLI command that replaced it, and the ones that moved to skills-optional/ — is removedIn2_0 in plugins/specweave/marketplace.json.


Built With SpecWeave

12 production projects shipped in 3 months. 5 in the App Store.

AppPlatformWhat It Does
EasyChampWeb (GCP)Enterprise sports league management. 20+ microservices, ML video analytics. 4 years in production.
SketchMateApp StoreAI drawing game — multi-model evaluation judges player art semantically.
LullaApp StoreBaby sleep app with Apple Watch. ML cry classification (tired/hungry/pain).
Football 2026App Store + WebWorld Cup 2026 companion. AI travel planner, live tickets, team stats.
SkillUp FootballApp StoreCoaches monetize training via Stripe. Instagram-like feed, scheduling.
BizZoneApp StoreStudent & business events with AI-powered news generation.
EduFeedWebNotebookLM meets Zoom. Upload videos, get quizzes, flashcards, live rooms.
JobWeaveWebAI-powered job search. Smart matching, resume optimization.
SpecWeavenpmThe framework itself. 600+ increments, 538+ releases.
SpecWeave UmbrellaGitHubMulti-repo orchestration workspace for all repositories.
vskillnpmPackage manager for AI skills. Security scanning, 49 platforms.
verified-skill.comWebSkill marketplace & studio. 105K+ verified skills, eval system.

Browse increments on GitHub — full transparency.


How It Compares

CapabilityCursor RulesCopilot InstructionsWindsurfClineVibe CodingSpecWeave
Structured specs (Problem, ACs, Approach)Yes
One closure gate you can actually see (verify.json)Yes
Autonomous execution (hours, unattended)Yes
Multi-agent teams (parallel, contract-first)Yes
External sync (GitHub / JIRA / ADO)Yes
Append-only ledger (claims, evidence, no lost work)Yes
LSP code intelligence (198x faster)Yes
Cross-tool handoff (any vendor, any subscription)Yes

Cursor tells AI "use Tailwind." SpecWeave tells AI "build a checkout flow against these five acceptance criteria, prove the tests pass, review the diff, then close."


Key Features

Spec-First Planning — Every feature starts as spec.md (Problem, Scope, ACs, Approach) plus tasks.md. Configuration, not prompting.

Evidence, not vibesspecweave task done --run "<test>" refuses a failing command and stores the exit code and output tail in the ledger.

Multi-agent, any vendor — A worktree per agent, claims through ledger.jsonl, one closure. Coordination happens only through committed files.

┌──────────────────┬──────────────────┬──────────────────┐
│ Agent 1 (auth) │ Agent 2 (payments)│ Agent 3 (catalog)│
│ T-01..T-04 │ T-05..T-08 │ T-09..T-12 │
│ ████████░░ 80% │ ██████░░░░ 60% │ ████░░░░░░ 40% │
└──────────────────┴──────────────────┴──────────────────┘

LSP Code Intelligence — 198x faster than grep, 0 false positives. Semantic references, definitions, and types.

10 skills, not 51 — see The ten skills above. Plus five standalone skills for non-Claude tools.

External Syncspecweave sync push|pull|status|setup. GitHub is first-class; Jira and Azure DevOps are opt-in.

Enterprise Ready — Compliance audit trails. Brownfield analysis. Multi-repo workspaces.

Dashboard — Built-in web dashboard for increment progress, analytics, cost tracking, and multi-project monitoring.


Skills Ecosystem

SpecWeave skills are published and verified at verified-skill.com. The vskill package manager provides:

  • Security scanning — 52 attack patterns, SHA-256 pinning, blocklist API
  • 49 agent platforms — one install deploys to Claude Code, Cursor, Copilot, Windsurf, and 45 more
  • Skill evals — unit tests, A/B comparisons, cross-model testing. Skills tested like programs.
  • Visual Skill Studiovskill eval serve for benchmarks, comparisons, and history
npx vskill install remotion-best-practices    # Install from registry
npx vskill eval run my-skill # Run eval suite

Documentation

spec-weave.comSpecWeave 2.0 · commands · skills · configuration

Contributing

Inside this repo dependency install scripts are disabled (.npmrc): run npm ci, then npm run setup (rebuilds the allowlisted native deps), and npm run security:scan before pushing — see SECURITY.md.

Community

Discord · YouTube · GitHub Issues

License

MIT — github.com/anton-abyzov/specweave