Graft

Infrastructure as Code for Claude Code multi-agent pipelines.

Write .gft files — or just describe what you want in natural language — and Graft compiles them into Claude Code harness structures with compile-time token budget analysis, runtime quality validation, and automatic fix suggestions.

Best for: Natural-language I/O pipelines — code review, ideation, content generation, data analysis, debate architectures. In coding workflows, use Graft for the NL sub-steps (review, analysis, planning) while running code execution directly.

77% fewer tokens, 200x faster than manual configuration. Benchmarks

Getting Started (2 minutes)

npm install -g @jsleekr/graft    # Install (Node.js 20+)
graft init my-project            # Scaffold project
cd my-project
claude                           # Open Claude Code — it already knows .gft

Then just say:

“Make a code review pipeline with security, logic, and performance reviewers running in parallel, then a senior reviewer.”

Claude Code writes the .gft file, compiles it, done. No DSL learning required — graft init injects the full .gft syntax spec into .claude/CLAUDE.md.

Or Write .gft Yourself

context PullRequest(max_tokens: 2k) {
  diff: String
  description: String
}

node SecurityReviewer(model: sonnet, budget: 4k/2k) {
  reads: [PullRequest]
  produces SecurityAnalysis {
    vulnerabilities: List<String>
    risk_level: String
  }
}

node SeniorReviewer(model: opus, budget: 6k/3k) {
  reads: [SecurityAnalysis, PullRequest]
  produces FinalReview {
    approved: Bool
    summary: String
  }
}

edge SecurityReviewer -> SeniorReviewer | select(vulnerabilities, risk_level) | compact

graph Review(input: PullRequest, output: FinalReview, budget: 25k) {
  SecurityReviewer -> SeniorReviewer -> done
}

graft compile review.gft    # Generate .claude/ harness
graft run review.gft --dry-run  # Validate with quality checks

The Closed Loop

Describe what you want (natural language)
  → Claude Code writes .gft
  → graft compile → graft run
  → Formatted results + quality validation
  → Automatic fix suggestions
  → Apply fixes → rerun

After execution, graft run automatically:

1. Formats results — node table with model, timing, token usage bar
2. Validates quality — schema, type, range, empty field, and budget checks
3. Suggests fixes — concrete .gft modifications for each issue

── Quality Check ─────────────────────────────────
  ✓ findings OK [Analyzer.findings]
  ⚠ Field 'recommendations' is empty [Writer.recommendations]
  ✓ Token budget OK: 74%
────────────────────────────────────────────────
Quality: 75% (3/4 checks passed)

── Suggestions ───────────────────────────────────
  ⚠ Field 'recommendations' is empty.
    → Increase Writer output budget: budget: 10k/10k
────────────────────────────────────────────────

Why Graft?

	Manual (natural language)	Graft (.gft)
Total tokens	~3,162	~716 (77% less)
Files to manage	~9	1
Time	~30 seconds	~148ms (200x faster)
Quality validation	None	Automatic
Consistency guarantee	None	Compiler-enforced

How It Works

.gft Source  →  Graft Compiler  →  .claude/ output  →  Claude Code runs it
                    ↓
            Scope + Type + Budget analysis (compile-time)

Graft Source	Generated Output	Purpose
node	.claude/agents/*.md	Agent with model, tools, output schema
edge \| transform	.claude/hooks/*.js	Data transform between nodes
graph	.claude/orchestration.md	Step-by-step orchestration plan
memory	.graft/memory/*.json	Persistent state across runs
config	.claude/settings.json	Model routing, budget, hook registration

Documentation

• Playground — try .gft in your browser with live compilation
• User Guide — full walkthrough of the language, CLI, and workflow
• Examples — 8 runnable .gft pipelines
• Benchmarks — NL vs .gft comparison data
• Language Specification — formal grammar and semantics
• npm package — @jsleekr/graft
• GitHub — source code

Using Graft in Coding Workflows

Graft handles NL sub-steps within a larger coding workflow:

Manual: Plan → Code (direct) → graft run review.gft → Fix (direct) → Done
                                       ↑
                              Graft handles this part

Use Graft for review, analysis, planning, and ideation steps. Run code execution through direct Claude Code sessions.

CLI

graft init [name]              # New project, or add Graft to existing project
graft compile <file.gft>      # Compile to .claude/ harness
graft check <file.gft>        # Parse + analyze only
graft import [dir] [-o file]  # Reverse-compile .claude/ into .gft
graft run <file.gft>          # Compile, execute, validate, suggest fixes
graft test <file.gft>         # Test with mock data
graft fmt <file.gft> [-w]     # Format .gft source
graft generate <desc>         # Generate .gft from natural language
graft watch <file.gft>        # Watch and recompile on changes
graft visualize <file.gft>    # Pipeline DAG as Mermaid diagram

License

MIT