菜单
Simulate-Deploy-Augment loop replacing traditional PDCA
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Automatically extract reusable patterns from Claude Code sessions and save them as learned skills for future use.
Instinct-based learning system that observes sessions via hooks, creates atomic instincts with confidence scoring, and evolves them into skills/commands/agents.
Research-before-coding workflow. Search for existing tools, libraries, and patterns before writing custom code. Invokes the researcher agent.
MECE+ methodology for breaking problems into independently-solvable atomic tasks
Recursive sandboxing for 100+ iterations before human review
Patterns and utilities for building React components in house-maint-ai
| name | SDA Controller |
| description | Simulate-Deploy-Augment loop replacing traditional PDCA |
Replaces traditional PDCA (Plan-Do-Check-Act) with a faster, AI-native iteration loop.
| PDCA | SDA | Improvement |
|---|---|---|
| Plan | Simulate | Test before building |
| Do | Deploy | Canary first |
| Check | Augment | Real-time learning |
| Act | (Continuous) | No manual cycle |
┌─────────────────────────────────────────────┐
│ │
│ SIMULATE ──► DEPLOY ──► AUGMENT │
│ ▲ │ │
│ └───────────────────────┘ │
│ │
└─────────────────────────────────────────────┘
Run solution in high-fidelity sandbox before any real deployment.
simulate:
environment: sandbox
actions:
- Run test suite: `npm run test`
- Build check: `npm run build`
- Load test: `npm run load-test:smoke`
success_criteria:
tests_pass: true
build_success: true
performance_threshold: "p95 < 200ms"
Small-scale canary release with rollback capability.
deploy:
strategy: canary
initial_traffic: 5%
ramp_schedule:
- 5%: 10min
- 25%: 30min
- 50%: 1hr
- 100%: 2hr
rollback_trigger:
error_rate: "> 1%"
latency_p99: "> 500ms"
Agents learn from telemetry, adjusting in real-time.
augment:
telemetry_sources:
- Error logs
- Performance metrics
- User feedback signals
feedback_actions:
- Auto-fix common patterns
- Generate hotfix PRs
- Update configuration
learning_loop:
- Store successful patterns
- Avoid failed approaches
/iterate simulate # Run simulation only
/iterate deploy # Deploy with canary
/iterate full # Complete SDA cycle
/iterate rollback # Emergency rollback
| Component | Purpose |
|---|---|
eval-harness | Powers SIMULATE phase |
verification-loop | Continuous checks |
| GitHub Actions | DEPLOY automation |
| Sentry | AUGMENT telemetry |
| Metric | Target |
|---|---|
| Simulation pass rate | 100% |
| Canary success rate | > 99% |
| Mean time to augment | < 5min |
| Rollback frequency | < 1% |