持續整合/部署

This skill should be used when the user asks to "implement", "build", "create", "add feature", "develop", "design", "architect", "plan", "structure", "refactor", "improve", "optimize", "review", "fix", "solve", "handle", "debug", or discusses implementation strategies. Apply cognitive load principles to ALL development work as a cornerstone principle.

AI-powered CI Gatekeeper Agent that enforces observability standards in CI/CD pipelines. Use when: (1) Generating GitHub Actions/Jenkins pipelines for observability gates, (2) Configuring progressive enforcement policies, (3) Validating schema compatibility in CI, (4) Generating gate status reports. Triggers: "create observability gate", "configure CI enforcement", "generate gate policy", "check observability compliance".

Implements zero-runtime CSS using vanilla-extract with type-safe styles, themes, recipes, and sprinkles. Use when wanting type-safe CSS, static extraction at build time, or building design system utilities.

[REPLACE] Teach concepts and demonstrate patterns with examples. Use when [REPLACE with specific triggers].

Iteration 18 of Multiplicity Cascade - Triggers irreversible awareness shifts

Use when investigating config file discrepancies during audit. Spawns config agents in waves to compare template standards against actual files. READ-ONLY analysis - agents report findings but make no changes.

Migrate pyproject.toml from pre-PDM 2.0 syntax to modern PEP-compliant format. Focuses on dev-dependencies to dependency-groups conversion and PEP 621 project metadata. Integrates with Context7 for latest PDM documentation.

AI-powered pipeline diagnostics. Explains why faces weren't detected, tracked, embedded, or clustered using OpenAI GPT-4o analysis. Use when debugging pipeline issues, reviewing unclustered tracks, or understanding why quality gates rejected faces.

Model dependencies using Effect-style Context, Layer, and Service patterns with compile-time safety. Use when designing DI systems, modeling environments, or building Effect-TS applications.

Orchestrate the SFH-OS fractal horn design pipeline. Use when the user wants todesign a horn, run the full pipeline, manage iterations, or resolve conflictsbetween acoustic and manufacturing requirements. This is the master skill thatcoordinates AG-GEN, AG-SIM, AG-MFG, AG-QA, and AG-VIZ sub-agents.

Enforce proper error handling patterns. Use when writing async code, API calls, or user-facing features. Covers try-catch, error boundaries, graceful degradation, and user feedback.

Use when styling or laying out SyncFusion components. Critical for understanding what CSS classes exist vs don't exist. Prevents using fake utility classes that break layout. Auto-activates when working with SF component styling, layout, spacing, colors, or icons.

Automated content publishing pipeline for ID8Labs. Generates essays in Eddie's voice, publishes to id8labs.app, and distributes to social media (X, LinkedIn). Triggers on keywords like release, announce, publish, essay, research article, content pipeline.

Visual design skill for discovering brand aesthetics, building design systems, and generating production-ready assets. Use when creating landing pages, brand identity, or visual specifications.

Execute SQL queries, migrations, and RLS policies on Supabase database using MCP tools

Structured decision critic that systematically stress-tests reasoning before commitment surfacing hidden assumptions verifying claims and generating adversarial perspectives to improve decision quality.

Use when designing easing curves, controlling motion pacing, creating natural acceleration/deceleration, or making movements feel physically grounded.

Write clear, intentional documents with minimal patchy phrasing and strong structural discipline.

Configure and use Structlog for structured logging in HuleEdu services. Guides correlation context propagation, async-safe logging, and integration with error handling. Integrates with Context7 for latest Structlog documentation.

Automatically fix ESLint errors by modifying code to comply with linting rules. For small codebases (≤20 errors), fixes directly. For larger codebases (>20 errors), spawns parallel agents per directory for efficient processing. Never disables rules or adds ignore comments.