// Pragmatic coding standards: concise, self-documenting code. 20 lines/function, 3 args max. Use when enforcing clean code, naming conventions, SRP, DRY, or KISS principles. NOT for architecture decisions (use system-design) or testing patterns (use test-architect).
AI-optimized browser automation with context-efficient @ref element handles. Use when automating browser interactions for AI agents with minimal context consumption. NOT for manual E2E testing (use e2e-automation) or screenshots (use chrome-devtools).
Prompt engineering for AI image generation: Midjourney, DALL-E, Stable Diffusion, Flux. Use when writing image prompts, configuring generation parameters, or optimizing AI art output. NOT for UI design decisions (use design-system) or media file processing (use media-processing).
API design patterns for REST, GraphQL, and tRPC services. Use when designing endpoints, versioning APIs, or choosing API architecture. NOT for code implementation (use nodejs-pro/python-pro) or database schema (use data-modeler).
Authentication and authorization: OAuth2, JWT, RBAC/ABAC, MFA, Passkeys, session management. Use when implementing login, permissions, tokens, or security flows. NOT for vulnerability scanning (use security-scanner) or API design (use api-architect).
Puppeteer CLI scripts for browser screenshots, debugging, and Core Web Vitals measurement. Use when capturing screenshots, measuring performance, or debugging browser issues. NOT for E2E test suites (use e2e-automation) or AI browser automation (use agent-browser).
Production deployment workflows, CI/CD pipelines, rollback strategies, and release management. Use when deploying apps, configuring pipelines, or planning rollback procedures. NOT for git operations (use git-workflow) or Vercel-specific deploys (use vercel-deploy).
| name | code-craft |
| description | Pragmatic coding standards: concise, self-documenting code. 20 lines/function, 3 args max. Use when enforcing clean code, naming conventions, SRP, DRY, or KISS principles. NOT for architecture decisions (use system-design) or testing patterns (use test-architect). |
| metadata | {"author":"pikakit","version":"3.9.206","category":"code-quality","triggers":["code style","clean code","best practices","naming conventions","SRP","DRY","KISS"],"coordinates_with":["code-review","test-architect","code-constitution"],"success_metrics":["0 functions > 20 lines","0 nesting > 2 levels","0 magic numbers"]} |
Concise. Direct. Measurable. Fixed thresholds: 20 lines/function, 3 args, 2 nesting levels.
Required: None — Code Craft is a knowledge-based skill with no external dependencies.
| Situation | Action |
|---|---|
| Writing production code | Apply core principles + naming/function rules |
| Reviewing code quality | Run full-review check |
| Editing existing files | Check dependency impact first |
| Completing any task | Run 4-item self-check |
| Architecture review | Read rules/engineering-spec.md |
| Owned by This Skill | NOT Owned |
|---|---|
| Naming conventions (4 rules) | Lint/format execution (→ code-review) |
| Function design (5 rules with thresholds) | Test writing (→ test-architect) |
| Code structure (4 patterns) | Constitutional governance (→ code-constitution) |
| Dependency awareness protocol | Language-specific style guides |
| Pre-completion self-check (4 items) | Automated code modification |
Pure decision skill: Produces coding guidance and evaluation results. Zero side effects.
| Principle | Rule |
|---|---|
| SRP | Each function/class does ONE thing |
| DRY | Extract duplicated code; reuse |
| KISS | Simplest solution that works |
| YAGNI | Don't build features until needed |
| Boy Scout | Leave code cleaner than you found it |
| Element | Convention |
|---|---|
| Variables | Reveal intent: userCount not n |
| Functions | Verb + noun: getUserById() not user() |
| Booleans | Question form: isActive, hasPermission, canEdit |
| Constants | SCREAMING_SNAKE: MAX_RETRY_COUNT |
| Rule | Threshold |
|---|---|
| Size | Max 20 lines; target 5–10 |
| Responsibility | Does one thing only |
| Abstraction | One level of abstraction per function |
| Arguments | Max 3; prefer 0–2 |
| Side Effects | Don't mutate inputs unexpectedly |
| Pattern | Rule |
|---|---|
| Guard Clauses | Early returns for edge cases |
| Flat > Nested | Max 2 levels of nesting |
| Composition | Small functions composed together |
| Colocation | Keep related code close |
| Check | Why |
|---|---|
| What imports this file? | Dependents might break |
| What does this file import? | Interface changes propagate |
| What tests cover this? | Tests might fail |
🔴 Rule: Edit the file + ALL dependent files in the SAME task.
| # | Check | Question |
|---|---|---|
| 1 | ✅ Goal met? | Did I do exactly what user asked? |
| 2 | ✅ Files complete? | Did I modify all necessary files? |
| 3 | ✅ Code works? | Did I verify the change? |
| 4 | ✅ No errors? | Lint and TypeScript pass? |
🔴 Rule: If ANY check fails, fix it before completing.
| Code | Recoverable | Trigger |
|---|---|---|
ERR_INVALID_REQUEST_TYPE | No | Request type not supported |
ERR_MISSING_CODE | Yes | Code snippet not provided |
ERR_MISSING_FILE_PATH | Yes | File path required for dependency check |
ERR_EMPTY_CODE | Yes | Code snippet is empty |
WARN_UNKNOWN_LANGUAGE | Yes | Language not recognized; generic rules applied |
Zero internal retries. Deterministic evaluation; same code = same violations.
| ❌ Don't | ✅ Do |
|---|---|
| Comment every line | Delete obvious comments |
| Helper for one-liner | Inline the code |
| Deep nesting (3+ levels) | Guard clauses; max 2 levels |
| Magic numbers | Named constants |
| God functions (50+ lines) | Split by responsibility; max 20 lines |
| File | Description | When to Read |
|---|---|---|
| verification-scripts.md | Validation scripts | Running automated checks |
| engineering-spec.md | Full engineering spec: contracts, security, scalability | Architecture review |
| Item | Type | Purpose |
|---|---|---|
code-review | Skill | Code quality review and lint execution |
test-architect | Skill | Test writing and coverage |
code-constitution | Skill | Constitutional governance |
⚡ PikaKit v3.9.206
