Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/rcsnyder-design
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: design
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/design
What's inside
- Skill name
design- Description
- Design phase. Takes scope.md and produces design.md with architecture, directory structure, interfaces, integration details, and justified complexity exceptions.
- License
- MIT
- Instructions
- 116 lines in SKILL.md · view on GitHub
Files (1)
- SKILL.md
SKILL.md (first lines)
Read `scaffolding/scope.md`. Produce `scaffolding/design.md`. ## Steps 1. Read `scaffolding/scope.md` fully 2. Read `preferences.md` if it exists (for stack conventions) 3. Produce `scaffolding/design.md` with these exact sections: ### design.md format ````markdown # Design: [Project Name] ## Architecture [How the pieces fit together. Include a simple ASCII diagram if the system has more than 2 components.] ## Directory Structure [The actual file tree you'll create at the repo root. Be specific.] ``` ├── src/ │ └── ... ├── tests/ ├── scaffolding/ ├── README.md └── ... ``` ## Interfaces [Key data shapes, API contracts, module boundaries. At least one concrete type/shape.] ## External Integrations [What this talks to outside itself. For each: - What it is - How you call it - What happens when it fails - **Test strategy**: mock (fake responses) / recorded (record-replay) / live (real API calls in tests)] [Or "None — self-contained" if applicable.] ## Observability [What needs logging, monitoring, or tracing. What would you check first if this breaks at 2am? - **Shed**: Structured logging to stdout. Errors include context (what input caused it, what state was expected). - **House**: Structured logging to Loki + Grafana alerting on error patterns. OpenTelemetry traces for request flows. Correlation IDs. Health endpoint. - **Skyscraper**: OpenTelemetry instrumentation (traces + metrics + logs). Prometheus + Loki + Grafana dashboards. Alerting on SLO breach.] ## Complexity Exceptions [Any justified place where BUILD may need to exceed the normal slice/file limits, share a cross-cutting abstraction, or stage work across multiple `AC-*` items. If none, write "None."] ## Open Questions [Anything uncertain. Resolve these before building. Or "None — straightforward."] ```` 4. **Design Review** (house/skyscraper only — skip for sheds): Walk through 2-3 key scenarios from the acceptance criteria against the design. For each: - Trace the data/control flow through the architecture - Identify where it could fail or degrade - Note any concerns (critical / major / minor) If critical issues are found, fix the design before proceeding. Append findings to design.md under a `## Design Review` section. 5. Run the **post-design gate**: - [ ] `scaffolding/design.md` exists - [ ] Has "Directory Structure" section - [ ] Has "Interfaces" section with at least one data shape - [ ] Every external integration has error handling noted - [ ] Every external integration has a test strategy declared (mock / recorded / live) - [ ] Has "Observability" section - [ ] Has "Complexity Exceptions" section (it may say `None.`) - [ ] No open questions remain unresolved (or explicitly deferred with rationale) - [ ] Design review completed (house/skyscraper) or skipped (shed) with rationale 6. If any gate condition fails, fix it and recheck. 7. Log the result to `scaffolding/log.md`: ```markdown ## DESIGN — [timestamp] - **Gate**: PASS (attempt N) - **Evidence**: [what was checked — directory structure, interfaces, integration test strategies] - **Changes**: scaffolding/design.md created - **Retries**: [total gate attempts this phase] - **Next**: ANALYZE ``` 8. Git checkpoint: ``` git add -A && git commit -m "docs(design): architecture for [project]" -m "[summarize architecture, key interfaces, integration count]\nGate: post-design PASS (attempt N)." ``` 9. **Auto-continue to ANALYZE** (unless user specified stepped mode). ``` ✓ DESIGN complete. Gate passed. Ready for ANALYZE. Continue? ``` ## Rules - Design should be pragmatic, not theoretical - Match the scale to the project — don't over-architect a shed - If an `AC-*` appears to require unusual complexity, record that explicitly under `## Complexity Exceptions` so BUILD does not discover it accidentally. - If scope.md says "GitHub Pages," don't design a Kubernetes deployment
From RCSnyder/lights-out-swe-plugin, MIT.
More from RCSnyder/lights-out-swe-plugin
- 0.474score
distill
RCSnyder
Distill raw input materials into structured reference docs.
SkillCFreshness: CClaude Code +9 - 0.471score
audit-stack
RCSnyder
Audit preferences.md stack choices against input docs for orthodox, idiomatic fit.
SkillCFreshness: CClaude Code +9 - 0.471score
build
RCSnyder
Build phase.
SkillCFreshness: CClaude Code +9 - 0.471score
build-discipline
RCSnyder
Execute BUILD, ITERATE, or verify-fix work in small, test-first vertical slices.
SkillCFreshness: CClaude Code +9 - 0.471score
analyze
RCSnyder
Analyze phase.
SkillCFreshness: CClaude Code +9 - 0.471score
init
RCSnyder
Scaffold the Lights Out SWE harness (copilot-instructions.md, preferences.md, LIGHTS-OUT-SWE.md, scaffolding/, docs/input/, .gitignore) into the current repo.
SkillCFreshness: CClaude Code +9 - 0.471score
iterate
RCSnyder
Iterate on a shipped project.
SkillCFreshness: CClaude Code +9 - 0.471score
review
RCSnyder
Review phase.
SkillCFreshness: CClaude Code +9 - 0.471score
expand
RCSnyder
Start a new project.
SkillCFreshness: CClaude Code +9 - 0.469score
deploy
RCSnyder
Deploy phase.
SkillCFreshness: CClaude Code +9 - 0.469score
verify
RCSnyder
Verify phase.
SkillCFreshness: CClaude Code +9 - 0.467score
reconcile
RCSnyder
Reconcile scaffolding documents against the codebase.
SkillCFreshness: CClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- May 24, 2026 (5 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from RCSnyder
- 0.529score
uefn-verse
RCSnyder
Implement maintainable Verse for a small UEFN gameplay slice.
SkillBFreshness: BClaude Code +9 - 0.526score
uefn-slice
RCSnyder
Convert a loose Fortnite/UEFN gameplay request into one tiny vertical slice with explicit acceptance criteria and scope boundaries.
SkillBFreshness: BClaude Code +9 - 0.526score
uefn-mcp
RCSnyder
Use UEFN's current Unreal MCP toolsets safely and efficiently.
SkillBFreshness: BClaude Code +9 - 0.526score
uefn-playtest-feedback
RCSnyder
Convert short human UEFN playtest feedback into one falsifiable gameplay hypothesis and the next tiny vertical slice.
SkillBFreshness: BClaude Code +9 - 0.526score
uefn-runtime-proof
RCSnyder
Verify a UEFN gameplay slice with compile, session, runtime, and log evidence while clearly separating machine-provable behavior from human playtest judgment.
SkillBFreshness: BClaude Code +9 - 0.524score
uefn-setup
RCSnyder
Set up or repair the connection between UEFN and an MCP-compatible coding agent.
SkillBFreshness: BClaude Code +9
Cite this page
design skill. Agent Mart. https://agentmart.dev/skills/rcsnyder-design (last updated 2026-10-24)