RohiRIK/skills：面向 AI 智能体的技能库

AdaptiveExplainers/ Choose and verify text, diagrams, interactive HTML, or video; generated HTML defaults to ~/Downloads. Agy/ Delegate a coding, refactor, PR-review, or image task to the Antigravity CLI (agy) as an autonomous worker, then monitor and report. Art/ Generate images, diagrams, and visual output. Use for any visual artifact. BackendDesign/ Reference for API design, database schema, and server architecture. Build/ Implement plan tasks one at a time with compile and commit gates. CodingStandards/ Language coding standards — TS, Python, Bash, PowerShell, Swift, Rust. CodebaseOnboarding/ Analyze an unfamiliar repo → architecture map, entry points, conventions, starter CLAUDE.md. ContextBudget/ Audit Claude Code context-window consumption and produce prioritized token-savings recommendations. CreateMcp/ Build a Model Context Protocol (MCP) server with the official SDK — tools, resources, prompts, transport, security. Scaffold, connect, debug. CreatePlugin/ Package a project as one shared AI-agent plugin with thin Hermes, Claude, OpenClaw, Pi, OpenCode, and MCP adapters. CreateCLI/ Generate a production-ready TypeScript CLI (3 tiers: parseArgs / Commander / oclif), Bun-only, gated by Verify. CreateCLI-Agent/ Generate a production-ready TypeScript CLI for autonomous agents: JSON contract, non-interactive, semantic exits. CreateCLI-Human/ Generate a production-ready TypeScript CLI for human operators: TTY UX, prose help, optional JSON. CreateHook/ Generate lifecycle hooks for AI agent runtimes — Claude Code, OpenCode, OpenClaw, Hermes Agent. CreateSkill/ Build and maintain skills in canonical structure. DataReportBuilder/ Turn a raw Excel/CSV dataset into a two-layer stakeholder report — plain-language Summary + untouched Raw Data. bun + ExcelJS engine. DockerPatterns/ Reference for Docker and Docker Compose local-dev patterns. Email/ Write, reply to, rewrite, and localize emails in English and Hebrew: brief-first workflows, 23 native bilingual templates, tone and culture rules, an AI-tell and calque lint, RTL-safe HTML rendering. Drafts only, never sends. FrontendDesign/ Reference for React and Next.js component design patterns. FrontendAesthetics/ Anti-slop pipeline — DesignBrief before code, SlopAudit gate after, Redesign protocol; craft library of font pairings, OKLCH palettes, banned tells. GitWorkflow/ Reference for git branching, commit conventions, merge vs rebase, and conflict resolution. GitHubOps/ Manage a GitHub repo via the gh CLI — hygiene, changelog, commit/push, PRs, releases, branch cleanup. Hygiene/ Audit ~/.claude for git, skill, code, and rules hygiene issues. Iterate/ Bounded PLAN→ACT→VERIFY→REFLECT iteration toward a goal, with a state file and hard exit conditions (/iterate). MacPluginPlaybook/ Phased playbook to build or port a native SwiftUI macOS menu bar app: research, scaffold, test-first logic, screenshot-driven UI, settings + onboarding, audit, PR. Includes lessons from a real build. OpenCode/ Delegate a coding/refactor/PR-review task to the OpenCode CLI as an autonomous worker. Orchestrate/ Decompose a spec into a dependency DAG, run units in parallel via delegation, review each in a separate context. Pi/ Delegate a coding/refactor/PR-review task to the Pi (pi.dev) CLI as an autonomous worker. Prompting/ Vendor-agnostic prompt-engineering standard library for authoring prompts, skills, agents, rules. Reflect/ Self-rate the just-finished output on five axes with evidence, then fix small gaps in place. Research/ Research a question at three depths (quick / standard / deep) by delegating to Agy/OpenCode/Pi + web search, then synthesize cited findings. SecurityReview/ Audit code for vulnerabilities and secrets. SkillForge/ Audit the whole skill library for agentic readiness and instrument it with telemetry in bulk. Simplify/ Post-implementation dead-code cleanup (/simplify). Spec/ Explore code + memory, then write acceptance criteria before planning. StrategicCompact/ Reference for context-compaction strategy and timing. Test/ Single TDD entry point for test-first features and bug fixes; runs FeatureTdd or ProveIt. Verify/ Six-phase quality gate (build→type→lint→test→secret→diff) ending in a READY / NOT READY verdict (/verify). Workflows/ Index of composed skill chains — which skills to run, in what order, for a full job. Routes a task to the right chain; full steps in workflows/.