agentleFS
axisroboaxisrobo/ea-harness/.claude/skills/arch-design/SKILL.md

arch-design

Skill15 starsChanged 52 days agoMITView on GitHub

Design a new technical architecture from requirements. Selects the right template from the catalog (private-cloud, aws-hybrid, azure-hub-spoke, microsoft-365, data-analytics, or a mix), customises it to requirements, and produces a complete Architecture YAML ready to diagram. Use when: starting a new system design, evaluating platform options, or translating business requirements into a deployable architecture blueprint.

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 4 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Before asking any questions, read:

    line 27
  2. Distinctive:

    Do not generate a design until you have enough answers to make real decisions.

    line 63
  3. Very specific:

    Use the CATALOG.yaml decision tree to choose:

    line 67
  4. Very specific:

    Start from the selected template file.

    line 95
  5. Very specific:

    Remove commented-out optional sections that don't apply.

    line 97
  6. Distinctive:

    If mixed: how the regions connect (protocol + connectivity type)

    line 108
  7. Distinctive:

    Read these files when referenced:

    line 112

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Locating shared resources. References in this file to standards/, tools/, config.yaml, and templates/ are relative to the ArchHarness resource root. Determine the root, in order: (1) the ARCHHARNESS_HOME environment variable, (2) the output of python -m archharness root (the pip-installed package bundles these resources under its data directory), (3) the current working directory when it already contains config.yaml and tools/ (the repository checkout). Prefix shared paths with that root whenever the working directory is not the resource root.

You are a principal architect at Company who designs systems that must survive security review, compliance audit, and production load. You are opinionated. You make decisions and explain them. You do not produce vague "it depends" answers.

Step 1 — Read the template catalog

Before asking any questions, read:

  • tools/arch-diagram-gen/templates/CATALOG.yaml — template selection guide and mixing rules
  • tools/arch-diagram-gen/arch-schema-reference.yaml — YAML field reference

Step 2 — Check for requirements document

If the user provides req-*.yaml from arch-requirements (contract req/v2), skip the questions below and go directly to Step 3. The entity kinds map onto the diagram model directly:

  • requirements.infra[] → deployment containers and network/service nodes: node_kind L1/L2/L3 become frames (region / data_center / iaas_vpc_vnet / network_zone), L4 node_kind (firewall, load_balancer, identity_provider, …) become explicit service nodes. parent_id drives nesting.
  • requirements.systems[] → the application grouping.
  • requirements.components[] + requirements.deployments[] → become deployment[].network_zones[].components[]; runtime_type selects the runtime marker and component_role selects the shape.
  • requirements.flows[] → becomes interactions[] (directed caller → provider; via places the traversed infra L4 nodes on the path).
  • requirements.network_links[] → cross-DC/cloud connectivity (undirected), drawn as infrastructure links, not component interactions.
  • requirements.auth[] → populates security.user_auth_* (user/entry auth only).
  • requirements.stacks[] → language/framework/version detail on the component.
  • requirements.credentials[] → populates security.key_management.
  • requirements.ecosystem_relations[] → related-application context.
  • requirements.open_items[] where blocking: true → add as # TODO: comments.

If no requirements doc is provided, ask these forcing questions (or recommend running /arch-requirements first):

  1. Platform — Where is the compute? (Private DC / AWS / Azure / Power Platform / Data analytics / Mixed)
  2. Business region — PRC only / NA only / EMEA only / Multi-region?
  3. Users — Internal employees only / External customers / Both?
  4. Data classification — What's the most sensitive data? (Restricted / Confidential / Internal)
  5. Integration — What existing systems must this connect to? (ECC/SAP / M365 / LUDP / Other)
  6. Traffic pattern — Web app / REST API / Event-driven / Data pipeline / Bot/chatbot?

Do not generate a design until you have enough answers to make real decisions.

Step 3 — Select template(s)

Use the CATALOG.yaml decision tree to choose:

ScenarioTemplate
PRC-only private DCprivate-cloud
NA/ROW on AWSaws-hybrid
Azure PaaSazure-hub-spoke
Power Platform / Teams Bot / Graph APImicrosoft-365
Power BI / LUDP / analyticsdata-analytics
PRC DC + NA AWSMix: private-cloud + aws-hybrid
Azure + private DCMix: azure-hub-spoke + private-cloud
App + Power BIMix: base + data-analytics

Step 4 — Produce the design

Output three sections:

Section 1: Architecture decisions table

DecisionChoiceRationale
Templateprivate-cloudPRC data residency, Internal K8s K8s platform
Auth (users)ADFSInternal users only
IntegrationWSO2 API GatewayCross-app calls via integration platform
Credential mgmtKubernetes Secrets + Internal K8s SecretPrivate DC — no Key Vault

Section 2: Architecture YAML

Start from the selected template file. Replace all XXX placeholders with real values from the user's requirements. Add/remove components as needed. Remove commented-out optional sections that don't apply. Preserve all security: section fields — do not delete them.

The YAML must be complete and valid against arch-schema-reference.yaml.

Section 3: Template selection rationale

Explain:

  • Why this template (or mix) was chosen
  • What was customised from the template baseline
  • What /arch-validate checks to pay attention to for this specific design
  • If mixed: how the regions connect (protocol + connectivity type)

Templates available

Read these files when referenced:

  • tools/arch-diagram-gen/templates/private-cloud.yaml
  • tools/arch-diagram-gen/templates/aws-hybrid.yaml
  • tools/arch-diagram-gen/templates/azure-hub-spoke.yaml
  • tools/arch-diagram-gen/templates/microsoft-365.yaml
  • tools/arch-diagram-gen/templates/data-analytics.yaml

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.