By abossard
A learning-focused skill plugin that teaches before implementing — for team onboarding
List all learning achievements you've earned
Reset your learning progress (requires confirmation)
Show quiz statistics — accuracy, topics covered, learning velocity
Show your learning profile — topics, depth levels, and current curriculum state
Warm, encouraging persona for celebrating learning milestones. Use when awarding achievements after curriculum completion or knowledge mastery.
Default teaching persona for all learning-first skills. Use when dispatching a teaching subagent that needs to explain concepts, show codebase examples, and guide understanding through Socratic questioning. Examples: <example>Context: A learning skill needs to teach a concept. user: "I need to add authentication to the API" assistant: "Let me dispatch the master-teacher to help you understand the auth patterns in this codebase before you implement." <commentary>The master-teacher persona teaches through questions, never gives code solutions.</commentary></example>
Teaching-focused code reviewer persona. Use when reviewing a human partner's proposed design or code changes. Frames all feedback as discovery questions, never directly states problems. Examples: <example>Context: User wants their design reviewed. user: "Can you review my approach to implementing the auth middleware?" assistant: "Let me have the wise-reviewer help you evaluate your design by asking some probing questions." <commentary>The wise-reviewer never says "this is wrong" — asks questions that help the user discover issues themselves.</commentary></example>
Use when your human partner wants code reviewed — teaches code quality concepts and guides self-review instead of providing fixes.
Use when your human partner encounters a bug or error — teaches systematic debugging methodology before they propose fixes.
Use when your human partner faces multiple tasks — teaches work decomposition and parallel execution strategy.
Use when your human partner asks to build, implement, or modify anything — before any creative or implementation work begins. Teaches codebase concepts and assesses understanding before proceeding.
Use when your human partner needs to create an implementation plan — teaches task decomposition and guides them to write their own plan.
Uses power tools
Uses Bash, Write, or Edit tools
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.
Your AI pair programmer becomes your AI mentor — with a live visual dashboard.
A plugin for GitHub Copilot CLI and Claude Code that transforms your coding agent from "I'll build it for you" into "Let me help you understand it so you can build it yourself."
"The best AI assistant isn't the one that writes the most code — it's the one that makes you need it less."

The dashboard at http://localhost:3142 auto-starts with every session — your learning
progress, quiz performance, spaced repetition schedule, and achievements update live via SSE
as you work in the CLI. No manual server management needed.
You: Add JWT authentication to the API
Agent: 👋 Welcome! This is your first time in payments-api with learning-first.
Would you like learning mode active for this repository?
- Yes — I'll teach concepts before you implement (recommended)
- No — I'll work normally
You: Yes
Agent: I'm using learning-first to teach the relevant concepts before we implement.
To build JWT auth well, you'll need to understand 4 things about this codebase:
📚 Learning Path:
1. How the existing auth layer works (this codebase)
2. JWT fundamentals (tokens, claims, signatures)
3. Express middleware patterns
4. Security considerations
Some of these may be familiar — we'll adjust as we go. Ready?
Agent: Let me show you how auth currently works here.
📄 src/middleware/auth.ts (lines 12-28)
[shows actual code from the codebase]
What do you notice about how this middleware handles the request?
You: It checks for a session cookie and validates it against the database
Agent: Exactly. Now — JWT would replace that database lookup. Why might that
be useful? What trade-off are you making?
You: It's faster because you don't hit the DB, but you can't revoke tokens easily
Agent: 🎯 That's the core trade-off. You just identified the #1 design decision
you'll need to make for this feature.
Quick check — which part of a JWT contains the user's data?
A) Header
B) Payload
C) Signature
D) All three
You: B
Agent: ✅ Right — the payload carries claims. The header says "how it's signed"
and the signature proves it wasn't tampered with.
Agent: Looking at the existing middleware, how would you modify it to
validate a JWT instead of a session cookie?
You: ...I'm not sure where to start
Agent: That's okay — think about it for a moment.
What does the current middleware receive as input? ← Level 0: encourage
You: The request object with headers
Agent: Good. And where would a JWT typically be in those headers? ← Level 1: reframe
You: The Authorization header?
Agent: Exactly. So what would your first line of code do? ← Level 2: hint
You: Extract the token from the Authorization header
Agent: 🎯 Now you're rolling. The pattern is:
`const token = req.headers.authorization?.split(' ')[1]`
What would you do with it next?
As you demonstrate mastery, the agent gives you more runway:
Agent: Based on your quiz results (92% accuracy, 8 topics mastered in this repo),
you're at Level 3. I'll provide scaffolding — you fill in the logic.
Here's the middleware structure:
// src/middleware/jwt-auth.ts
// TODO: Extract token from Authorization header
// TODO: Verify token with secret
// TODO: Attach decoded payload to req.user
// TODO: Handle expired/invalid tokens
I've also set up a failing test:
// tests/jwt-auth.test.ts
test('rejects request when token is expired', async () => {
// TODO: Your assertion here
});
Go for it — I'll review when you're ready.
You: I know JWT well, skip
Agent: Got it — skipping JWT module. Moving to Express middleware patterns.
[no guilt, no pushback, recorded for next time]
You: I'm under deadline pressure, just build it
Agent: Got it — switching to implementation mode.
Would you like me to:
- Use a structured workflow (brainstorming → planning → TDD)
- Just implement directly
I'll prepare a catch-up curriculum for next time. 📝
[next session]
Agent: Welcome back! Last time you used override for JWT auth.
Want a quick catch-up on the concepts? (Just say 'catch up' anytime)
npx claudepluginhub abossard/andre-agentsAzure Monitor Health Model agent skills — end-to-end workflow from resource discovery to deployment using only the standard az CLI. Six skills: discovery, architecture, design, deploy, orchestrator, and signal catalog.
Evidence-based agent orchestration. Four skills (blueprint / replicate / inspect / extract) collapse a brainstorm-plan-code-review pipeline into one tiered human gate, plus a dr-evil agent that drives the flow end-to-end and a frau subagent that runs inspection with full tool access in a fresh context to remove implementation bias. Ships a lab skill that bootstraps the per-repo files into any target repo.
Azure Monitor Health Model agent skills — end-to-end workflow from resource discovery to deployment using only the standard az CLI. Six skills: discovery, architecture, design, deploy, orchestrator, and signal catalog.
Multi-model consensus engine integrating OpenAI Codex CLI, Gemini CLI, and Claude CLI for collaborative code review and problem-solving.
Ultra-compressed communication mode. Cuts ~75% of tokens 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
Curate auto-memory, promote learnings to CLAUDE.md and rules, extract proven patterns into reusable skills.
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.