What it does
X-Ray reads a codebase and extracts its architecture without an LLM, so the code stays on the local machine. Its outputs include an ARCHITECTURE.md map with a module dependency graph, domain model, routes with auth status and risk findings; an AGENT_CONTEXT.md file for rules files; a machine-readable architecture.json; and Mermaid diagrams with an HTML viewer. It also gives a 0 to 100 production-readiness score across eight dimensions, with evidence and the top fixes ranked by points lost. Findings cover hardcoded secrets, raw SQL that bypasses the ORM, auth tokens in browser storage and routes without auth guards.
It suits reviewing AI-generated code before it ships, auditing an inherited codebase, or tracking whether a repository drifts from production readiness over time, since the same repo produces the same score. An agent can call tools such as blast_radius to see what depends on a file, arch_summary to locate routes and services, boundary_findings, production_readiness, generate_diagrams and xray_scan to re-scan after structural changes.
The quickest setup is the Claude Desktop extension, a bundle that prompts for the repository to analyse. The server is also listed in the Official MCP Registry under its name, so compatible clients can install it from there. Extraction covers Python (Flask, FastAPI, SQLAlchemy, Django, Celery) and JavaScript/TypeScript (Next.js, Express, Prisma). It uses only the standard library and makes no network calls.
Limitations and requirements
Extraction is limited to the Python and JavaScript/TypeScript stacks listed above. Auth status is reported as unknown when a guard is attached at router level and cannot be detected per function. The production-readiness score is not a security audit, according to the publisher. Claude Desktop installation requires selecting the repository to analyse.
Safety
Our team has not reviewed this resource. Only automated rules checked it; resources with high-risk signals are not published. These are signals, not a guarantee that installing it is safe. Try it in a sandbox project first.
- Runs code or changes an assistant's behaviour
Sources
- Official MCP Registry first seen 13 Nov 2026 · last seen 13 Nov 2026 · its page on the source
io.github.Anioko/archiet-audit
Submit architecture docs, get a consulting-grade traceability audit at a shareable URL
v0.1.2As shown on registry.modelcontextprotocol.io when we last read it