Developer Toolkit
5 files that turn Claude Code into an engineering lead. Container-ready code structure, AI-powered compliance checking, and a full development workflow — all from one prompt.
Free. Works with any language.
What You Get
How It Works
Copy the toolkit files into your project
cp ~/Downloads/anchorscape-developer-toolkit/* your-project/All 5 files go in your project root. Works with new or existing projects.
Open Claude Code and paste the bootstrap prompt
Run claude in your project directory, then copy the entire contents of CLAUDE_BOOTSTRAP_ANCHORSCAPE_V2.md as your first message.
Answer Claude's questions
Claude will assess your project, propose a tech stack (or analyze existing code), and ask for approval before proceeding.
Claude sets everything up
ANCHORSCAPE-STANDARD.md ← The 10 rules (stays in root)
TECH_STACK_DECISION.md ← Tech choices + rationale
.claude/agents/ ← 5 sub-agents (agent + template moved here)
src/ ← Scaffolded to standard
tests/
Dockerfile ← Multi-stage, non-root
.dockerignore
.env.example
For existing projects: Claude runs the AnchorScape Agent first, generates a compliance report with a score, fixes critical issues, then switches to the development workflow.
Use Without Anchorscape
Everything works standalone. Your code will be container-ready and deployable to any platform:
When you're ready for automated scanning, one-click fixing, and managed deployment — that's where Anchorscape comes in.
FAQ
Does this work with any language?
Yes. The bootstrap adapts to your tech stack — Node.js, Python, Go, Rust, Java, or anything else. The AnchorScape Standard is language-agnostic.
What if I already have a CLAUDE.md?
It gets backed up to CLAUDE.md.backup before the new one is created. Existing project-specific rules are merged in.
Can I use just the Standard without the bootstrap?
Absolutely. The ANCHORSCAPE-STANDARD.md and ANCHORSCAPE-AGENT.md work independently. Tell any AI tool to "read ANCHORSCAPE-AGENT.md and act as that agent" and it'll analyze your codebase.
What does uploading my report do?
It creates a scan report in Anchorscape with your score and findings — same as if we scanned it ourselves. You can then use Shipwright to auto-fix issues or deploy with pre-configured settings. Saves time and API costs.
Does this work with Cursor, Windsurf, or Copilot?
The AnchorScape Agent file works with any AI coding tool. The bootstrap prompt is optimized for Claude Code's sub-agent system, but the Standard and Agent files are universal.
Want the Full Platform?
The toolkit gets your code container-ready. Anchorscape takes it further: AI-powered scanning that finds what rules can't, one-click auto-fixing, and managed Kubernetes deployment.