Borrowing it
Nothing to install: this file belongs to cognitx-leyton/codegraph. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/cognitx-leyton/codegraph/main/.claude/commands/arch-check.mdgit clone --depth 1 https://github.com/cognitx-leyton/codegraphWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/cognitx-leyton/codegraph/arch-check)<a href="https://agentmods.dev/commands/cognitx-leyton/codegraph/arch-check"><img src="https://agentmods.dev/badge/commands/cognitx-leyton/codegraph/arch-check/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/cognitx-leyton/codegraph/arch-check"><img src="https://agentmods.dev/badge/commands/cognitx-leyton/codegraph/arch-check.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00016 | $0.01038 |
| Opus 5.5 | $0.00006 | $0.00415 |
| Sonnet 5.5 | $0.00003 | $0.00208 |
| Haiku 4.5 | $0.00002 | $0.00104 |
Grade A, and why
arch-check scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 83 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Usage
/arch-check
No args. Runs three built-in policies and reports violations. Zero violations = architecture still holds. Non-zero = investigate.
| Policy | What it detects |
|---|---|
| Import cycles | File IMPORTS paths of length 2-6 that loop back to the start |
| Cross-package imports | twenty-front importing from twenty-server (front/back boundary) |
| Controller → Repository bypass | Controllers that call Repository methods within 3 hops without going through a *Service class — skips the service layer |
Each policy is an independent Cypher block. To extend the default policy set, edit this file and add a new block following the pattern.
What this does
echo "=== Policy 1: file-level IMPORTS cycles ==="
codegraph query "
MATCH path = (a:File)-[:IMPORTS*2..6]->(a)
RETURN [n IN nodes(path) | n.path] AS cycle,
length(path) AS hops
ORDER BY hops ASC
LIMIT 10
"
echo ""
echo "=== Policy 2: cross-package boundary (twenty-front -> twenty-server) ==="
codegraph query "
MATCH (a:File)-[:IMPORTS]->(b:File)
WHERE a.package = 'twenty-front' AND b.package = 'twenty-server'
RETURN a.path AS importer, b.path AS importee
LIMIT 10
"
echo ""
echo "=== Policy 3: Controller -> Repository bypass (skips *Service) ==="
codegraph query "
MATCH (ctrl:Controller)-[:HAS_METHOD]->(m:Method)-[:CALLS*1..3]->(target:Method)
MATCH (repo:Class)-[:HAS_METHOD]->(target)
WHERE repo.name ENDS WITH 'Repository'
AND NOT EXISTS {
MATCH (ctrl)-[:HAS_METHOD]->(:Method)-[:CALLS*1..3]->(:Method)<-[:HAS_METHOD]-(svc:Class)
WHERE svc.name ENDS WITH 'Service'
}
RETURN ctrl.name AS controller, repo.name AS repository, target.name AS method
LIMIT 10
"
Caveats
- Policies are project-specific. The default set is tuned for this repo (codegraph) + Twenty. A fresh repo with different conventions will see either false positives or zero coverage. Rewrite the policies in this file — it's designed to be forked.
- Cycle bound is
*2..6. Very long cycles (depth > 6) won't be detected. If you suspect one, raise the upper bound (slower query). - The Controller→Repository check is a heuristic. It relies on naming convention (
*Service,*Repository). Projects that use different names need to edit theENDS WITHclauses. - Python coverage is limited. No
:Controllerlabel for Python yet (NestJS-only in Stage 1). Policy 3 is effectively TS-only.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2mo ago First seen · 83 lines · 16 tokens per session scan A 9231fd4f40fe
arch-check is a command published in the GitHub repository cognitx-leyton/codegraph (12 stars, last pushed 3mo ago), licensed Apache-2.0. It adds 16 tokens to every session and 1,038 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other commands, from other repositories
grasp-build-graph
Run graspanalyze + graspminimalcontext for the current repo, store the sessionid, and report a 100-token orientation.
feedback
A design-stage security review for product requirement documents and other planning documents; a PRD is a document describing what a product should do.
va
A read-only security review of a codebase across eight architecture areas, with findings checked against file-and-line evidence.
compliance
A follow-up security check that verifies whether earlier findings were fixed and, depending on the selected mode, looks for new problems.
pentest
A read-only penetration-testing command for examining source code from an attacker’s perspective and producing proof-of-concept checks. Penetration testing is authorized security testing that tries to verify exploitable weaknesses.
run
A security-workflow coordinator that turns a natural-language request into a plan involving code review, verification, attack testing, or reporting.