Software Architecture Maestro

Production-grade software and system architecture Skill for AI agents.

software-architecture-maestro designs, reviews, audits, and modernizes full-code, low-code, no-code, SaaS-configured, workflow-automation, RPA, composable, hybrid, distributed, cloud, data-intensive, AI-native, mobile, edge, IoT, platform, and internal systems.

It does not jump from a request to a fashionable stack. It runs a traceable architecture process:

Request
-> Intake and evidence
-> Diagnosis
-> Context and workload classification
-> Architecture drivers
-> Delivery modality by capability
-> Domain and team boundaries
-> Option generation
-> Economic and technical trade-offs
-> System design
-> Failure and scenario testing
-> Companion Skill routing
-> Implementation specification
-> Fitness functions
-> Governance
-> Quality review

What makes it different

• Treats architecture as a sociotechnical decision system.
• Separates facts, constraints, assumptions, hypotheses, recommendations, and vendor claims.
• Selects full-code, low-code, no-code, SaaS, automation, RPA, composable, hybrid, or manual delivery per capability.
• Connects architecture to team ownership and cognitive load.
• Requires systems of record, data ownership, failure behavior, recovery, security, observability, economics, migration, and exit design.
• Produces ADRs, C4 guidance, implementation contracts, scenario tests, and architecture fitness functions.
• Blocks high readiness when critical gaps remain.
• Recommends independent Airtable, ClickUp, and n8n companion Skills when tool-specific implementation is selected.

Coverage

Core architecture

Application architecture, modular monoliths, microservices, distributed systems, event-driven systems, domain-driven design, data architecture, integration, capacity, performance, reliability, resilience, security, observability, deployment, cloud, platform engineering, multi-tenancy, and architecture governance.

Delivery modalities

Manual processes, configured SaaS, no-code, workflow automation, RPA, low-code, composable architectures, hybrid systems, and full-code systems.

Specialized systems

AI-native and agentic systems, mobile and offline-first applications, edge systems, IoT and cyber-physical systems, global and regional systems, internal operations platforms, marketplaces, financial ledgers, and multi-tenant SaaS.

Evolution and assurance

Legacy modernization, migration, transition architecture, build-buy-compose decisions, architecture economics, TCO, vendor lock-in, exit strategy, fitness functions, architecture tests, audits, anti-patterns, scenario testing, ADRs, technology radar, and implementation handoff.

Repository structure

software-architecture-maestro-repo/
├── software-architecture-maestro/
│   ├── SKILL.md
│   ├── agents/openai.yaml
│   ├── references/
│   ├── scripts/
│   └── assets/
│       ├── schemas/
│       ├── templates/
│       ├── rubrics/
│       └── taxonomies/
├── examples/
├── tests/
├── docs/
├── tools/
├── .github/workflows/
├── dist/skill.zip
├── README.md
├── CONTRIBUTING.md
├── SECURITY.md
├── CHANGELOG.md
├── LICENSE
└── requirements-dev.txt

Install

Upload dist/skill.zip to an environment that supports custom Skills. The installable root is software-architecture-maestro/ and the only entrypoint is SKILL.md.

Validate locally

python -m pip install -r requirements-dev.txt
python software-architecture-maestro/scripts/validate_artifact.py examples/hybrid-vendor-operations/architecture-intake.json
python software-architecture-maestro/scripts/validate_architecture_package.py examples/hybrid-vendor-operations
python software-architecture-maestro/scripts/audit_architecture_repository.py software-architecture-maestro
python -m unittest discover -s tests -v

Route an architecture request

python software-architecture-maestro/scripts/route_architecture_request.py \
  --request "Design a hybrid vendor operations system with Airtable, ClickUp, and n8n" \
  --output architecture-routing.json

Companion Skills

This repository does not bundle or depend on tool-specific Skills. When the architecture selects those platforms, it can recommend:

• he8um/airtable-skills
• he8um/clickup-skills
• he8um/n8n-skills

Installation requires explicit user permission and a supported installation action. Each companion remains independently installable and independently maintained.

Safety

Do not commit secrets, production exports, private infrastructure details, customer data, or credentials. Architecture generation does not authorize infrastructure provisioning, deployment, migration, purchasing, or production changes.

License

MIT