AI agent skill · by @FlorianBruniaux
Self Assessment
Interactive skill assessment with personalized learning path generation
Archived 2026-10-23 · Source updated 2026-10-10 · License: CC-BY-SA-4.0. Check the original repository for the latest version.
SKILL.md instructions
# Codex Self-Assessment **Master your Codex skills with a personalized learning journey.** This interactive assessment identifies your current expertise level and generates a custom learning path tailored to your gaps and goals. --- ## How It Works ### Step 1: Choose Your Mode **Quick Mode** (5 min) - 10 questions across 10 topics - Best for: Quick skill check, understanding your general level - Output: Overall profile (Beginner/Intermediate/Advanced) + top 2-3 gaps **Comprehensive Mode** (20 min) - 20 questions (2 per topic) - Best for: Detailed mastery assessment, planning team training - Output: Per-topic scores + detailed learning path + 3 practice projects ### Step 2: Answer Questions Questions cover: - Practical usage (how to do things) - Decision-making (when to use what) - Architecture patterns (why things work) - Production best practices (security, performance, scaling) Each question has: - 4 options (A-D) - Immediate feedback on correct answer - Link to relevant guide section ### Step 3: Get Your Profile **Quick Mode Output:** ``` ╔═══════════════════════════════════════════════╗ ║ Your Profile: Intermediate Developer ║ ╠═══════════════════════════════════════════════╣ ║ Overall Score: 14/20 ║ ║ Level: Intermediate (strong) ║ ╠═══════════════════════════════════════════════╣ ║ Immediate Gaps: ║ ║ • MCP Servers (5/5 topics weak) ║ ║ • Advanced Patterns (4/5 weak) ║ ║ • Hooks (3/5 partially weak) ║ ║ ║ ║ → Recommended learning time: 4-6 hours ║ ╚═══════════════════════════════════════════════╝ ``` **Comprehensive Mode Output:** ``` ╔═══════════════════════════════════════════════════════════════╗ ║ Detailed Mastery Profile ║ ╠═══════════════════════════════════════════════════════════════╣ ║ ║ ║ 01 Quick Start ████░░░░░░ 8/10 Proficient ║ ║ 02 Core Concepts ██████░░░░ 6/10 Proficient ║ ║ 03 Memory & Settings ██████░░░░ 6/10 Proficient ║ ║ 04 Agents ████░░░░░░ 4/10 Basic ║ ║ 05 Skills ████░░░░░░ 4/10 Basic ║ ║ 06 Commands ██████░░░░ 6/10 Proficient ║ ║ 07 Hooks ██░░░░░░░░ 2/10 None ║ ║ 08 MCP Servers ██░░░░░░░░ 2/10 None ║ ║ 09 Advanced Patterns ░░░░░░░░░░ 0/10 None ║ ║ 10 Security ███░░░░░░░ 3/10 Basic ║ ║ ║ ║ OVERALL: Intermediate (14/100 points) ║ ╚═══════════════════════════════════════════════════════════════╝ ``` ### Step 4: Get Your Learning Path Based on your scores, you get: **Prioritized learning sequence** (respects dependencies): 1. Topics ordered by weakness + importance 2. Prerequisites shown (you can't skip foundations) 3. Time estimates for each topic (sums only what you need) 4. Templates and examples linked to each topic **Example path for Intermediate developer:** ``` Phase 1: Hooks Fundamentals (2 hours) → Read: guide/ultimate-guide.md#7-hooks → Try: examples/hooks/event-driven-formatting.sh → Quiz: /lesson-quiz 07-hooks Phase 2: MCP Integration (3 hours) → Read: guide/ultimate-guide.md#8-mcp-servers → Try: examples/mcp-configs/github-integration.json → Quiz: /lesson-quiz 08-mcp-servers Phase 3: Advanced Patterns (4 hours) → Read: guide/ultimate-guide.md#9-advanced-patterns → Practice: 3 real-world projects (see below) → Quiz: /lesson-quiz 09-advanced-patterns ``` **3 Practice Projects** Based on your profile, you get targeted exercises: - Project 1: Combine 2-3 topics you're weak in - Project 2: Real-world workflow integration - Project 3: Team/production scenario --- ## Getting Started ### Quick Assessment ``` /self-assessment quick ``` ### Full Assessment ``` /self-assessment comprehensive ``` ### Options ``` /self-assessment quick --verbose # Show explanations /self-assessment comprehensive --save # Save results to file ``` --- ## What You Learn This assessment is designed to identify: ### Fundamental Gaps - Core interaction loops you haven't tried - Essential features you've overlooked - Mental model misconceptions ### Decision-Making Blind Spots - When to use agents vs skills vs commands - Which hook event to use for which problem - MCP server selection for your needs ### Production Readiness - Security best practices - Performance optimization patterns - Scaling multi-agent systems - Team configuration and governance --- ## After Your Assessment ### Immediate Actions 1. **See your profile** — Understand your current mastery level 2. **Identify gaps** — Know exactly what to learn next 3. **Get a learning path** — Personalized, dependency-aware sequence 4. **Practice** — 3 targeted projects matching your level ### Recommended Follow-up Skills - `/lesson-quiz [topic]` — Quiz yourself on specific topics - `/audit-agents-skills` — Audit your own config quality - `/Codex-md` — Manage your AGENTS.md configuration ### Integration with Learning Path After assessment, the `/lesson-quiz` skill provides per-module verification: - Answer 5 questions per module - Get instant feedback on weak areas - Unlock next module when ready --- ## Scoring & Profiles ### Junior Developer (0-6 points) - Just installed Codex - Know basic commands and workflows - Haven't touched memory/agents/hooks yet - **Next step:** Core Concepts + Memory module ### Intermediate Developer (7-13 points) - Using Codex daily - Comfortable with AGENTS.md and basic agents - Exploring skills and commands - Some hook usage, no MCP servers - **Next step:** Deep MCP + Advanced Patterns ### Advanced Developer (14-20 points) - Master multi-agent orchestration - Production-grade security and scaling - Custom skills and complex hooks - MCP server expertise - **Next step:** Team training, enterprise patterns --- ## Tips for Better Results 1. **Answer honestly** — Don't guess, skip unknown questions 2. **Review explanations** — Each question has a guide link 3. **Revisit before level-up** — Retake assessment after learning new topics 4. **Share results** — Use your profile to coordinate with your team --- ## Need Help? - **Questions about the assessment?** → Check `/help self-assessment` - **Want to discuss results?** → See `guide/roles/learning-with-ai.md` - **Found an error?** → Report it on GitHub Issues --- **Ready to master Codex?** Start with `/self-assessment quick` now.