OpenClaude local LLM environment — oc-start launcher, settings, commands
Find a file
pookNast 2110c27c82 perf: optimize /rlm and /night-queue for token efficiency
rlm.md: 405→107 lines (14.6KB→4.7KB, -68%)
night-queue.md: 230→49 lines (7.5KB→2.6KB, -65%)
Combined: 22KB→7.3KB (-67% token savings per invocation)

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-04-24 21:40:33 -04:00
commands perf: optimize /rlm and /night-queue for token efficiency 2026-04-24 21:40:33 -04:00
config docs: add RLM engine reference with token optimization checklist 2026-04-24 19:44:06 -04:00
memory feat: replace with OpenClaude environment — skills, config, commands, memory 2026-04-24 19:05:52 -04:00
skills feat: replace with OpenClaude environment — skills, config, commands, memory 2026-04-24 19:05:52 -04:00
README.md feat: replace with OpenClaude environment — skills, config, commands, memory 2026-04-24 19:05:52 -04:00

OpenClaude Environment

Production-grade AI development environment — skills, hooks, commands, memory, and configuration for OpenClaude (Claude Code).

Structure

openclaude/
├── skills/                    # Claude Code skills (invoked via /skill-name)
│   ├── preflight/            # Masterclass pre-flight quality checklist (8 gates, 48 checks)
│   └── ui-ux-designer/       # WCAG 2.2 AAA UI/UX design skill
├── commands/                  # Slash command definitions
├── config/                    # Environment configuration (templates, redacted)
│   ├── CLAUDE.md             # Global Claude rules (Level 5000)
│   ├── CLAUDE_GLOBAL.md      # Private global instructions
│   ├── CLAUDE_PROJECT.md     # Project-level workspace map
│   ├── CLAUDE_WORKSPACE.md   # OpenClaude workspace rules + GLM-5.1 escalation
│   ├── mcp.json              # MCP server configuration
│   ├── settings.json.example # Settings template (secrets redacted)
│   └── .claude.json.example  # Claude config template (redacted)
├── memory/                    # Persistent memory (private)
│   ├── MEMORY.md             # Memory index
│   └── team/                 # Shared team memory
└── .gitignore

Skills

/preflight

Masterclass pre-flight quality checklist. 8 gates, 48 checks covering:

  1. Application Security (9 checks) — OWASP, injection, auth, rate limiting
  2. Cryptography & Data (5 checks) — TLS, encryption, sessions, key rotation
  3. Reliability (7 checks) — circuit breakers, idempotency, concurrency
  4. Performance (6 checks) — Core Web Vitals, bundles, caching, DB optimization
  5. Accessibility (6 checks) — WCAG 2.2 AA, keyboard nav, semantic HTML
  6. Observability (6 checks) — structured logging, RED metrics, health endpoints
  7. Deployment & CI/CD (7 checks) — pipelines, rollback, secrets, parity
  8. Compliance (5 checks) — GDPR, consent, DSAR, retention

Target: <1% post-ship remediation.

/ui-ux-designer

Senior lead product designer skill. WCAG 2.2 AAA, oklch() colors, Tailwind + shadcn/ui + HTMX + Alpine.js. 100-point checklist enforced.

Commands

Command Description
/finished Wrap up task — quality gates, summary, changelog, commit
/ralph-loop Generator-Evaluator agent loop with PRD tracking
/code-review Production code review
/level-5000-load-up Activate Level 5000 directives
/net-load Load homelab network context
/net-save Save network changes
/projects List projects from MEMORY.md

Setup

  1. Clone this repo
  2. Copy config/settings.json.example~/.claude/settings.json and fill in API keys
  3. Copy config/.claude.json.example~/.claude.json and fill in API keys
  4. Copy config/mcp.json~/.mcp.json
  5. Copy skills/~/.openclaude/skills/
  6. Copy commands/~/.openclaude/commands/
  7. Copy config/CLAUDE*.md to appropriate locations per CLAUDE.md workspace map

Dependencies

  • OpenClaude CLI
  • MCP Servers: homelab-mcp, context7, skill-seekers, tldr, token-savior
  • GLM-5.1 proxy (oc-proxy) for escalation tasks

License

Internal use — contains proprietary operational tooling.