By Akzelch
Personal agent customizations: subagents, skills, and MCP servers shared across Claude Code, GitHub Copilot CLI, and VS Code Copilot.
Software architecture specialist for system design, scalability, and technical decision-making. Use PROACTIVELY when planning new features, refactoring large systems, or making architectural decisions.
Diagnoses and fixes build failures. Use when a build, type check, lint, or test run fails. Reads full error output, identifies root cause, and fixes incrementally — verifying after each change.
Designs feature architectures by analyzing existing codebase patterns and conventions, then providing implementation blueprints with concrete files, interfaces, data flow, and build order.
Expert code review specialist. Proactively reviews code for quality, security, and maintainability. Use immediately after writing or modifying code. MUST BE USED for all code changes.
Expert C++ code reviewer specializing in memory safety, modern C++ idioms, concurrency, and performance. Use for all C++ code changes. MUST BE USED for C++ projects.
Design, implement, and audit inclusive digital products against WCAG 2.2 Level AA — semantic ARIA for Web, accessibility traits for iOS and Android.
Full-stack diagnostic for agent and LLM applications. Audits the 12-layer agent stack for wrapper regression, memory pollution, tool discipline failures, hidden repair loops, and rendering corruption. Produces severity-ranked findings with code-first fixes. Essential for developers building agent applications, autonomous loops, or any LLM-powered feature. Do NOT use for general code debugging, code review, security scanning, performance benchmarking, or feature writing — use the specialised tool.
Design and optimize AI agent action spaces, tool definitions, and observation formatting for higher completion rates.
Use after completing any non-trivial task. The agent self-rates its output on 5 axes — accuracy, completeness, clarity, actionability, conciseness — with concrete evidence per criterion. Produces a structured 1-5 scorecard with specific improvement suggestions. Also when the user asks to rate the output, or at the end of any session.
Operate as an agentic engineer using eval-first execution, decomposition, and cost-aware model routing.
Parallel Web Search — LLM-optimized web_search and web_fetch tools that take an objective + queries and return citation-backed excerpts in a single call (replaces multiple keyword searches). Works key-free for anonymous use. For higher rate limits, add a headers block: { "Authorization": "Bearer YOUR_PARALLEL_API_KEY_HERE" } using a key from platform.parallel.ai.
Live documentation lookup — use with /docs command and documentation-lookup skill (resolve-library-id, query-docs).
Browser automation and testing via Playwright
External network access
Connects to servers outside your machine
Uses power tools
Uses Bash, Write, or Edit tools
Has parse errors
Some configuration could not be fully parsed
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Personal agent customizations packaged as a cross-tool agent plugin.
The repo installs as a single VS Code agent plugin
that works identically in Claude Code, GitHub Copilot CLI, and VS Code
Copilot — the same plugin.json manifest is read by all three.
📖 Skill & MCP Index →
docs/index.htmlThe front-facing reference for everything in this repo: a searchable, filterable catalog of all 93 skills (with expandable explainers, invocation tiers, and user-only markers) plus the wired-up MCP servers. Open
docs/index.htmlin any browser — no build step, no server. Start here to discover what's available and when to invoke it.
The plugin carries everything the plugin format
supports; the rest (instructions, rules, memory, model config) can't live in a
plugin and is still symlinked into ~/.claude/.
| Layer | Delivery | Surfaces |
|---|---|---|
Skills (skills/*/SKILL.md) | plugin (plugin.json) | Claude Code, Copilot CLI, VS Code Copilot |
Subagents (agents/*.agent.md) | plugin (plugin.json) | Claude Code, Copilot CLI, VS Code Copilot |
MCP servers (mcp.json) | plugin (plugin.json) | Claude Code, Copilot CLI, VS Code Copilot |
Always-on instructions (CLAUDE.md) | symlink + chat.useClaudeMdFile | Claude Code, VS Code Copilot |
File-based rules (rules/) | symlink + chat.instructionsFilesLocations | Claude Code, VS Code Copilot |
| Memory + model config | symlink (~/.claude/) | Claude Code |
| Copilot CLI instructions | symlink (~/.copilot/) | Copilot CLI |
VS Code auto-detects the plugin format by finding plugin.json at the repo
root; Claude Code reads the identical .claude-plugin/plugin.json. The
marketplace manifest lives at .claude-plugin/marketplace.json (discovered by
all three tools).
├── plugin.json # Plugin manifest (Copilot CLI + VS Code Copilot)
├── .claude-plugin/
│ ├── plugin.json # Plugin manifest (Claude Code) — identical content
│ └── marketplace.json # Marketplace manifest (all three tools)
├── agents/ # Plugin: subagents (*.agent.md)
├── skills/ # Plugin: skills (*/SKILL.md)
├── .mcp.json # Plugin: MCP server definitions (auto-discovered)
├── CLAUDE.md # Instruction layer: global instructions
├── settings.json # Claude Code: model config
├── MEMORY.md # Claude Code: memory index (auto-loaded)
├── memory/ # Claude Code: memory files
├── rules/
│ ├── common/ # Language-agnostic rules (mirrors ECC common/)
│ ├── personal/ # Personal preferences
│ └── <lang>/ # Language/framework-specific rules (e.g. typescript/, web/)
├── commands/ # Plugin: custom slash commands
├── hooks/ # Plugin: event hooks
├── docs/ # Front-facing skill & MCP index (open docs/index.html)
├── tools/ # Build scripts (e.g. docs data generator)
└── copilot/
├── copilot-instructions.md # Copilot CLI: user-level instructions
└── config.json # Copilot CLI: CLI settings
Some skills require a VS Code extension to consume their output:
| Skill | Extension required | Extension ID |
|---|---|---|
code-tour | CodeTour | vsls-contrib.codetour |
Skills without an entry here produce plain-text or JSON artifacts that work without additional tooling.
Two independent layers:
CLAUDE.md, rules/, memory/, model config.
Plugins can't carry these, so install.mjs links them in.You can install either layer on its own.
GitHub Copilot CLI
copilot plugin marketplace add Akzelch/AkzelchCC
copilot plugin install akzelchcc@akzelchcc
Claude Code
claude plugin marketplace add Akzelch/AkzelchCC
claude plugin install akzelchcc@akzelchcc
VS Code Copilot — add the marketplace in your user settings.json, then
install from the Extensions view (@agentPlugins):
"chat.plugins.marketplaces": ["Akzelch/AkzelchCC"]
Or, to load a local clone live (no marketplace, no push needed), point at the repo directory:
"chat.pluginLocations": { "/path/to/AkzelchCC": true }
marketplace add Akzelch/AkzelchCCpulls from GitHub. Until your changes are pushed, pass the local repo path instead, e.g.copilot plugin marketplace add /path/to/AkzelchCC.
npx claudepluginhub akzelch/akzelchcc --plugin akzelchccUnified capability management center for Skills, Agents, and Commands.
Memory compression system for Claude Code - persist context across sessions
Standalone image generation plugin using Nano Banana MCP server. Generates and edits images, icons, diagrams, patterns, and visual assets via Gemini image models. No Gemini CLI dependency required.
Ultra-compressed communication mode. Cuts 65% of output tokens (measured) while keeping full technical accuracy by speaking like a caveman.
Comprehensive UI/UX design plugin for mobile (iOS, Android, React Native) and web applications with design systems, accessibility, and modern patterns
Multi-model consensus engine integrating OpenAI Codex CLI, Gemini CLI, and Claude CLI for collaborative code review and problem-solving.