| context | fork |
| user-invocable | false |
| name | quality-framework |
| description | ATLAS Quality Framework for Artibot: Automated, Tested, Learned, Adaptive, Secure. Defines the validation cycle, coverage targets, and GRPO-driven continuous quality improvement. Auto-activates when: reviewing code, setting quality gates, running a validation cycle, or assessing coverage. Triggers: quality framework, ATLAS, validation cycle, quality gate, coverage threshold, ํ์ง ๊ฒ์ดํธ, ์ปค๋ฒ๋ฆฌ์ง, ํ์ง ๊ฒํ . |
| lang | ["en","ko"] |
| level | 3 |
| triggers | ["quality","quality framework","ATLAS","validation cycle","์ปค๋ฒ๋ฆฌ์ง ์๊ณ๊ฐ","quality gate","์ฝ๋ ํ์ง ๊ฒํ ","quality standards","testing standards","security validation","performance standards","quality metrics","ํ์ง","ํ์ง ๊ฒ์ดํธ","์ปค๋ฒ๋ฆฌ์ง","ํ์ง ๊ฒํ "] |
| agent | Explore |
| allowed-tools | ["Read","Grep","Glob"] |
| agents | ["tdd-guide","security-reviewer","performance-engineer","refactor-cleaner"] |
| tokens | ~3K |
| category | code-quality |
| platforms | ["claude-code","gemini-cli","codex-cli","cursor"] |
| source_hash | c3db3f38 |
| whenNotToUse | Do not run the full 8-step ATLAS cycle on every micro-PR or hotfix. For single-file changes under 50 lines, run only the blocking gates (Steps 1, 2, 4, 5). Reserve full cycle execution for release candidates and architecture-level changes. |
ATLAS Quality Framework
Automated | Tested | Learned | Adaptive | Secure
Current State
!npm run lint 2>&1 | tail -3
!npm test 2>&1 | tail -5
When This Skill Applies
- Code review and quality assessment
- Defining acceptance criteria for features
- Setting up CI/CD quality gates
- Post-implementation validation
- Security and performance audits
- Coverage analysis and improvement
8-Step Validation Cycle
| Step | Gate | Type |
|---|
| 1 | Syntax โ linter, formatter | Blocking |
| 2 | Types โ tsc/mypy/cargo check | Blocking |
| 3 | Lint โ complexity โค10 cyclomatic | Warning |
| 4 | Security โ OWASP Top 10, secret scan | Blocking |
| 5 | Tests โ unit โฅ80%, integration โฅ70% | Blocking |
| 6 | Performance โ <200ms API, <3s page | Warning |
| 7 | Documentation โ public API, CHANGELOG | Warning |
| 8 | Integration โ no breaking changes, flags | Warning |
Coverage Targets
| Level | Target |
|---|
| Unit | โฅ 80% |
| Integration | โฅ 70% |
| Type coverage | 100% (no any) |
| Security | OWASP Top 10 |
Quick Reference
Blocking gates: Steps 1, 2, 4, 5 โ must pass to merge.
Warning gates: Steps 3, 6, 7, 8 โ flag for review, don't block.
Escalate immediately when:
- Any critical security vulnerability found (Step 4)
- Coverage drops >5% from baseline (Step 5)
- Performance regression >20% (Step 6)
Workflow checklist: See references/atlas-checklist.md
Output templates: See references/output-templates.md
Metrics dashboard: See references/metrics-dashboard.md
GRPO Quality Loop
After resolving any quality issue: observe context โ compare strategies โ rank by success rate โ update weights โ apply to future similar contexts. Enable via artibot.config.json quality pattern logging.
Common Rationalizations
- "ํ
์คํธ๋ ๋์ค์" โ ์๋ฐ: Step 5๋ Blocking gate
- "์ปค๋ฒ๋ฆฌ์ง 80%๋ฉด ์ถฉ๋ถํด" โ ๊ธฐ์ค์ ์ด์ง ๋ชฉํ๊ฐ ์๋; ํ๋ฝ ์ถ์ธ๊ฐ ๋ฌธ์
- "์ด๋ฒ PR์ ์์์ ATLAS ์ ํด๋ ๋ผ" โ hotfix ์ ์ธ ๋ชจ๋ PR์ ์ต์ Blocking gates ์ ์ฉ
- "๋ณด์ ์ด์๋ ๋ค์ ์คํ๋ฆฐํธ์" โ Step 4 critical = ์ฆ์ ์ฐจ๋จ, ๋ค์ ์คํ๋ฆฐํธ ์์
Red Flags
- lint/type ์ค๋ฅ ๋ฌด์ํ๊ณ ๋จธ์ง
- ์ปค๋ฒ๋ฆฌ์ง ์์น ์์ด "์ถฉ๋ถํ ํ
์คํธํ๋ค" ์ฃผ์ฅ
- security validation ๊ฑด๋๋ฐ๊ธฐ
- quality gate ์คํจ ํ ์ฌ์คํ ์์ด ์น์ธ ์์ฒญ