![]() |
Architectural Guardrail for AI-native development
CXGRD maps blast radius across your codebase, so AI tools know what they will break - before they write a line
CXGRD scans your project, builds a dependency graph, and gives AI assistants the architectural context they need to make safer changes.
- Tell you exactly what will break before a change lands.
- Enrich AI prompts with project structure, dependencies, and blast-radius analysis.
- Validate architecture with circular dependency checks, orphan detection, and compiler-backed validation.
- Enforce merge policies in GitHub Actions to protect your team from risky changes.
Typical workflow
# 1. Install npm install -g cxgrd # 2. Authenticate cxgrd auth login # Required for Pro, Team, and Enterprise plans # 3. Scan your project cxgrd scan # 4. Check blast radius before a change cxgrd input "extract UserService into a separate module" # 5. Generate an enriched prompt for your AI cxgrd prompt "extract UserService into a separate module" # 6. Validate the result cxgrd check
See it in action
CXGRD-Demo.mp4
Core commands
cxgrd scan→ writes .cg/ from scratch or diffs itcxgrd input→ reads graph + architecture + history, writes to history.jsoncxgrd prompt→ reads everything, sends a subgraph to your LLM, and returns an enriched promptcxgrd check→ reads graph + compiler output, writes result to history.jsoncxgrd auth login→ log in for Pro, Team, and Enterprise userscxgrd doctor→ verifies toolchain readiness before strict checkscxgrd watch→ monitors dependency changes in real timecxgrd init-hooks→ installs pre-commit checks automatically
For more information, visit the official website or read the docs here.

