archiet-audit

Archiet X-Ray

mcp-name: io.github.Anioko/archiet-xray

See what your AI sees. Your codebase is too big for any AI agent's context window. The agent reads 40 files out of 4,000, makes a change, and you have no way to know whether it respected the architecture — or quietly violated it.

X-Ray extracts the actual architecture of any repo — deterministically, no LLM, code never leaves your machine — and gives it to both you (an interactive map) and your AI agent (an MCP server + context pack).

$ python xray.py .

Archiet X-Ray v0.2.0 — your-repo
  visibility score : 78/100
  prod readiness   : 71/100 (near production-ready)
  code files       : 3,412 (1,907 read, 1,584 with elements)
  routes           : 214
  entities         : 87
  async tasks      : 31
  findings         : 6
  wrote            : .archiet/ARCHITECTURE.md
                     .archiet/AGENT_CONTEXT.md
                     .archiet/architecture.json
                     .archiet/diagrams/diagrams.html (+3 .mmd)

What you get

File	What it is
ARCHITECTURE.md	Human-readable map: module dependency graph (Mermaid), domain model, every route with auth status, dependency hotspots, risk findings, production-readiness scorecard
AGENT_CONTEXT.md	Drop into your CLAUDE.md / rules file — makes Claude Code, Cursor, and Windsurf respect your architecture today
architecture.json	The machine-readable model (the "repo genome")
diagrams/	Architecture diagrams as Mermaid sources + a self-contained HTML viewer — module dependency graph, ER / domain-model diagram, HTTP route map

Production-readiness score (new in 0.2)

A deterministic 0–100 production-readiness score over 8 dimensions: route auth coverage, secrets hygiene, client token storage, data-layer discipline, test footprint, migration discipline, ops readiness (Docker/CI/.env contract), and docs/API contract. Same repo always scores the same — so the score is comparable across commits, branches, and repos, and you can put it in CI.

Each dimension reports its evidence ("142/214 statically-guarded routes carry an auth guard") and the report ends with the top fixes ranked by points lost. Dimensions that don't apply (a CLI tool has no routes) take half credit and say so — honesty over flattery.

Useful when: reviewing AI-generated code before shipping it, auditing an inherited codebase, or tracking whether your repo is drifting away from production readiness over time.

Architecture diagrams from code (new in 0.2)

generate_diagrams turns any repo into Mermaid architecture diagrams, deterministically extracted from the code itself:

• Module dependency graph — the real import structure, not the wiki's
• ER / domain-model diagram — entities and relations from SQLAlchemy, Django, and Prisma models
• HTTP route map — the API surface grouped by prefix

You get the raw .mmd sources (paste into any README, wiki, GitHub, GitLab, Notion, Obsidian, or VS Code — they all render Mermaid natively) plus a self-contained diagrams.html viewer.

Give it to your agent (MCP)

# Claude Code (from PyPI — pip install archiet-xray)
claude mcp add archiet-xray -- archiet-xray mcp /path/to/repo

# or from a checkout
claude mcp add archiet-xray -- python /path/to/mcp_server.py /path/to/repo

Claude Desktop — download archiet-xray-<version>.mcpb from the latest release, then double-click it (or Settings → Extensions → Install Extension). It prompts for the repository to analyze; one-click after that.

Official MCP Registry — listed as io.github.Anioko/archiet-xray, so registry-aware clients can install it by name.

Your agent can now ask — before it edits:

• blast_radius — "who depends on this file? what breaks if I touch it?"
• arch_summary — "where do routes/entities/services actually live?"
• boundary_findings — "hardcoded secrets, raw SQL, tokens in localStorage, unauthenticated routes"
• production_readiness — "score this repo 0–100 for production readiness, with evidence and top fixes"
• generate_diagrams — "draw the module graph / ER diagram / route map as Mermaid"
• xray_scan — re-scan after structural changes

Principles

1. Deterministic. Same repo in, same map out. Every fact traces to a file and line. No LLM guesses anywhere in the pipeline.
2. Honest. What can't be extracted with confidence is labelled unmapped — never invented. A wrong map is worse than no map.
3. Local-first. Stdlib only, zero network calls, zero telemetry. Your code never leaves your machine.

What it extracts today

• Python: Flask / FastAPI routes (+ auth-guard detection), SQLAlchemy & Django models with relations, Celery tasks, import graph — via ast, not regex
• JS/TS: Next.js app & pages router (pages + API routes), Express routes, Prisma models, import graph
• Findings: hardcoded secrets, raw SQL bypassing the ORM, auth tokens in localStorage/AsyncStorage, routes without auth guards
• Graph: module dependency edges, fan-in/fan-out, blast-radius hotspots

More stacks (Go, Java, Rails, .NET) welcome — the extractor pattern is one class per language. PRs invited.

Real examples

examples/ holds unedited X-Ray output for repos you know — microblog (Flask), the official FastAPI full-stack template, and vercel/commerce (Next.js). GitHub renders the Mermaid maps inline. On the FastAPI template, X-Ray correctly detects CurrentUser dependency auth on 18 routes and flags a real auth-token-in-localStorage write in useAuth.ts.

FAQ

How is this different from a dependency-graph MCP server (Codegraph, dependency-mcp)? Those show call/import edges. X-Ray extracts web-architecture semantics on top of the graph: which routes exist, which carry auth guards, where the domain entities live, and where security boundaries leak (hardcoded secrets, raw SQL, tokens in localStorage). No graph tool tells you "510 of your routes have no detectable auth guard."

How is this different from a CLAUDE.md generator? CLAUDE.md generators write instructions and conventions — usually with an LLM. X-Ray extracts facts: every claim in its output traces to a file and line, and what it can't extract it labels unknown. Use both: your conventions plus X-Ray's ground truth.

What does auth status ? mean? "Not detectable from per-function analysis." FastAPI routers often attach auth at include_router(dependencies=...) level, which is invisible when analyzing the route function. X-Ray reports unknown rather than guessing a confident "no auth" — a wrong map is worse than no map.

Is the production-readiness score a security audit? No. It is a deterministic static-signal score — it measures whether the repo carries the marks of production discipline (auth guards on routes, no hardcoded secrets, tests, migrations, ops + docs contracts). The payload says so explicitly. It is a fast, repeatable triage number, not a substitute for code review, load testing, or a security audit.

Why Mermaid for the diagrams instead of images? Mermaid sources are diffable, render natively on GitHub/GitLab/Notion/VS Code/Obsidian, and stay in sync with the repo because you can regenerate them in one command. The bundled diagrams.html gives you a browser view without installing anything.

Does it phone home? No. Stdlib-only, zero network calls, zero telemetry. The only outbound anything is a link in the generated footer.

Sibling server: archiet-audit

X-Ray never touches the network. If you want the hosted counterpart — a procurement-grade architecture audit (traceability %, severity-ranked gaps, 30/60/90-day roadmap) generated from your architecture documents — the archiet-audit MCP server bridges your agent to archiet.com/audit-my-architecture. It is a separate server precisely so X-Ray's "your code never leaves your machine" guarantee stays absolute.

Privacy Policy

X-Ray is local-only by design. Specifically:

• Data collection: none. X-Ray collects no telemetry, no analytics, no usage data, and no personal information of any kind.
• Data usage: local analysis only. Your code is read from disk solely to build the architecture model, in-process, on your machine.
• Data storage: your disk, your control. The only artifacts produced are the files written to .archiet/ inside the repository you scan (ARCHITECTURE.md, AGENT_CONTEXT.md, architecture.json, diagrams/). Delete them at any time; nothing else is stored anywhere.
• Third-party sharing: none. X-Ray makes zero network calls (stdlib-only, no HTTP client is even imported). Your code never leaves your machine.
• Data retention: not applicable. No data reaches us, so there is nothing for us to retain or delete.
• Contact: [email protected] — or open an issue on this repo.

The company-wide policy at archiet.com/privacy covers Archiet's hosted products; it applies to X-Ray only in the trivial sense that X-Ray sends us nothing.

Part of Archiet

X-Ray is the free, open companion to Archiet — the architecture-to-code platform. The map X-Ray extracts is the same formal model Archiet uses to enforce architecture on every PR (boundary gates, drift scoring, consulting-grade architecture reports) and to regenerate production-ready applications from it.

MIT licensed.