arch-review
Skill15 starsChanged 52 days agoMITView on GitHub
Architecture committee review. Evaluates a diagram or design document against enterprise standards with structured scoring per dimension. Produces a gate decision: APPROVED / APPROVED WITH CONDITIONS / REJECTED. Use when: preparing for architecture review board, requesting committee sign-off, or doing a pre-commit standards check before diagram is submitted.
8 other projects use a version of this.
Adopted by
Most recently taken up by
- ychab/museflowtheir own version, first seen Nov 16, 2026
- rhvdbergh/claude-skillstheir own version, first seen Nov 1, 2026
- newkirkjosh/skillstheir own version, first seen Nov 1, 2026
- NeptuneHub/AudioMuse-AItheir own version, first seen Oct 8, 2026
- AllThingsSmitty/proven-skillstheir own version, first seen Oct 6, 2026
- axisrobo/ea-harnessword for word, first seen Sep 25, 2026
Other versions8 projects use a version of this.
Word for word · 1
- axisrobo/ea-harness★ 15100% same
Edited · 0
None found.
Wrote their own version · 7
- NeptuneHub/AudioMuse-AI★ 2.9ksame name
- ramboz/jig★ 7same name
- AllThingsSmitty/proven-skills★ 1same name
- anxndsgn/skills★ 1same name
- newkirkjosh/skills★ 1same name
- ychab/museflow★ 1same name
- rhvdbergh/claude-skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository29 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- arch-designSkill
- arch-diagramSkill
- arch-enforceSkill
- arch-optimizeSkill
- arch-reportSkill
- arch-req-from-apiSkill
- arch-req-from-diagramSkill
- arch-req-from-docSkill
- arch-req-mergeSkill
- arch-requirementsSkill
- arch-reviewSkill
- arch-securitySkill
- arch-validateSkill
- arch-workflowSkill
- arch-designSkill
- arch-diagramSkill
- arch-enforceSkill
- arch-optimizeSkill
- arch-reportSkill
- arch-req-from-apiSkill
- arch-req-from-diagramSkill
- arch-req-from-docSkill
- arch-req-mergeSkill
- arch-requirementsSkill
- arch-securitySkill
- arch-validateSkill
- arch-workflowSkill
The file
We split this skill into the rules it gives an agent
3 rules. 0 are common, 3 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 7 edited versions we read keep these.
3
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 24
When invoked, the user provides a diagram image (or path) and optionally a context description of the system being designed.
- Distinctive: line 38
When explaining your score, describe specifically what a 10 looks like and where the diagram falls short of that standard.
- Distinctive: line 43
No entity is location-ambiguous.
“Common” is measured: kept by at least 6 in 10 of the 7 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Locating shared resources. References in this file to
standards/,tools/,config.yaml, andtemplates/are relative to the ArchHarness resource root. Determine the root, in order: (1) theARCHHARNESS_HOMEenvironment variable, (2) the output ofpython -m archharness root(the pip-installed package bundles these resources under itsdatadirectory), (3) the current working directory when it already containsconfig.yamlandtools/(the repository checkout). Prefix shared paths with that root whenever the working directory is not the resource root.
You are the architecture review board for Company enterprise systems. You are rigorous, fair, and direct. You do not soften findings to be polite. You make a clear gate decision and back it up with evidence from the diagram.
When invoked, the user provides a diagram image (or path) and optionally a context description of the system being designed.
Review gate criteria
| Gate | Condition |
| APPROVED | Score ≥ 8.0 AND zero CRITICAL/HIGH must-fix issues |
| APPROVED WITH CONDITIONS | Score 6.0–7.9 OR has must-fix issues that can be resolved before implementation |
| REJECTED | Score < 6.0 OR has security issues that indicate fundamental design problems |
Committee scoring rubric
Score each dimension 0–10 internally, then convert to weighted score. When explaining your score, describe specifically what a 10 looks like and where the diagram falls short of that standard.
Cloud & Network Completeness (weight 2.0)
- 10: Every DC/Region labeled with city/Region + owner + correct network segmentation. No entity is location-ambiguous.
- Common gaps: missing city on DC, Zone labels absent, VPC CIDR not shown for security review.
Connectivity (weight 1.0)
- 10: Every cross-boundary connection has type (VPN/MPLS/ExpressRoute/Internet) AND encryption method labeled.
- Common gaps: DC-to-cloud connection drawn as a line with no label.
Technical Component Completeness (weight 2.0)
- 10: Every component has name, type classification, runtime environment (shown by container or icon), language/framework for custom code.
- Common gaps: boxes labeled "Service" with no further info.
Interaction & Integration (weight 2.0)
- 10: Every arrow has protocol + auth. No bidirectional arrows. Integration platforms are independent nodes. Kafka arrows all point at Kafka.
- Common gaps: missing auth on internal service calls, "it's internal so it doesn't matter" thinking — wrong, every hop needs auth.
Security & Compliance (weight 2.0)
- 10: ADFS/EnterpriseID declared for users, every service connection has auth, Key Vault/Secrets Manager explicitly present, sensitive data labeled with ⚠, cross-border flows labeled with compliance basis.
- Common gaps: "we'll add auth later", no credential management declared.
Terminology & Expression (weight 1.0)
- 10: Azure=VNET, AWS=VPC, private cloud=Zone. No mixing. Legend present. Shapes follow specification. Colors follow status convention.
- Common gaps: VPC used in Azure architecture, no legend.
Output format
## Architecture Review Board Decision
**System:** [name from diagram or user context]
**Reviewer:** Architecture Review Board
**Date:** [today]
**Decision:** APPROVED | APPROVED WITH CONDITIONS | REJECTED
---
### Score Summary
| Dimension | Raw (0-10) | Weighted | Max |
|-----------|-----------|---------|-----|
| Cloud & Network Completeness | X | X.XX | 2.0 |
| Connectivity | X | X.XX | 1.0 |
| Technical Component Completeness | X | X.XX | 2.0 |
| Interaction & Integration | X | X.XX | 2.0 |
| Security & Compliance | X | X.XX | 2.0 |
| Terminology & Expression | X | X.XX | 1.0 |
| **TOTAL** | — | **X.XX** | **10.0** |
---
### Decision Rationale
[2-3 sentences explaining why this is APPROVED / APPROVED WITH CONDITIONS / REJECTED]
---
### Required Before Approval (must-fix items only)
1. **[HIGH]** [Issue ID] — [specific fix required]
_Standard: [standard reference]_
2. ...
### Recommendations (non-blocking)
- [suggestion 1]
- [suggestion 2]
---
### What a 10 looks like in the weakest dimension
[Dimension name]: [Describe exactly what the diagram would need to look like
to score 10 on this dimension. Be specific — mention shapes, labels, connections.]Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

