sde-architecture

SDE Skills Marketplace

80+ SDE skills and 40+ chained workflows across 9 plugins — system design, code quality, architecture, debugging, execution, MERN stack, testing, DevOps, and career growth. JavaScript-first.

Designed for Claude Code and Cowork. Skills compatible with other AI assistants.

---

Start Here

Building a new system? -> /design Reviewing code or a PR? -> /review-code Writing an architecture decision? -> /write-adr Building a MERN feature? -> /mern-feature Debugging a production issue? -> /postmortem Breaking an epic into tickets? -> /write-tickets Prepping for an interview? -> /mock-design

---

Installation

Claude Code (CLI)

claude plugin marketplace add yourusername/sde-skills

Other AI assistants (skills only)

for plugin in sde-*/; do
  cp -r "$plugin/skills/"* ~/.gemini/skills/ 2>/dev/null
done

---

Plugins

Plugin	Skills	Commands	Domain
sde-system-design	8	4	HLD, estimation, data modeling, API design
sde-code-quality	8	4	PR review, refactoring, SOLID, patterns
sde-architecture	7	4	ADRs, migrations, resilience, observability
sde-debugging	6	4	RCA, postmortems, logs, incident response
sde-execution	7	5	Tickets, estimation, deployment, runbooks
mern-stack	12	6	MongoDB, Express, React, Node.js, JWT
sde-testing	6	3	Unit, integration, E2E, TDD, contracts
sde-devops	6	3	CI/CD, Docker, Kubernetes, Terraform
sde-career	8	5	Interviews, promotion, specs, feedback

---

Quick Reference

I want to...	Command
Design a system end-to-end	/design [system]
Estimate capacity and scale	/estimate [system]
Design a REST API contract	/design-api [resource]
Design a data model	/model-data [domain]
Review my PR before submitting	/review-code [diff]
Refactor messy code safely	/refactor [code]
Spot and fix code smells	/smell-check [code]
Plan tests for a feature	/test-strategy [feature]
Write an architecture decision	/write-adr [decision]
Plan a safe migration	/plan-migration [from to]
Write a blameless postmortem	/postmortem [incident]
Debug a production issue	/debug [symptom]
Respond to a live incident	/incident [description]
Break an epic into tickets	/write-tickets [epic]
Write release notes	/release-notes [commits]
Create a deployment checklist	/deploy-plan [service]
Write an on-call runbook	/write-runbook [service]
Build a full MERN feature	/mern-feature [feature]
Build JWT auth (MERN)	/mern-auth
Design a MongoDB schema	/mern-schema [domain]
Build an Express REST API	/mern-api [resource]
Build a React component	/mern-component [component]
Review MERN code	/mern-review [code]
Plan a full test suite	/test-plan [feature]
Generate tests for my code	/write-tests [code]
Design a CI/CD pipeline	/design-pipeline [service]
Dockerize a service	/dockerize [service]
Write Terraform IaC	/write-iac [infrastructure]
Practice system design	/mock-design [system]
Practice behavioral interview	/mock-behavioral
Write a promotion document	/write-promo [level]
Write a technical spec	/write-spec [feature]
Review my resume	/review-resume

---

Core Engineering Rules

These rules are encoded into every skill and command output.

1. Think Before You Code - Reason through requirements, constraints, and tradeoffs first
2. Production Mindset - Error handling, input validation, security, observability by default
3. Tradeoffs Are Mandatory - Why this approach, what it costs, alternatives considered
4. Quantify Everything - O(n^2) not "slow", "hits pool limit at 1K QPS" not "won't scale"
5. Security Is Non-Negotiable - Flag injection, secrets, missing auth, timing attacks
6. Explicit Over Clever - Readable names, justified complexity, comment intent not mechanics
7. No Hallucinated APIs - Note "verify syntax" when uncertain, use pseudocode
8. Fail Loudly, Recover Gracefully - Typed errors, circuit breakers, dead letter queues
9. Blameless by Default - Fix systems and processes, not people
10. Suggest the Next Step - After every command, offer the natural follow-on

---

License

MIT