BLOG: https://juntz-g1thub.github.io/#post=coding-skills-dev REPO: https://github.com/juntz-g1thub/CodingSkills.git

Background Recently I've been experimenting with vibe coding. I'm using OpenCode as my development tool. When developing TUI programs, AI-assisted code modification proved extremely difficult — every small bug fix required repeated manual testing, and feeding those results back to the agent was cumbersome. Especially for TUI-level feedback, I'd either take screenshots manually or type prompts. At one point, I nearly gave up on AI-assisted development.

Motivated by this, I started researching MCP tools. To help agents better utilize these MCP tools, I developed a set of skills combining one or more of them. These skills have been packaged and published to npm.

Released Skills 1. agc-debug — TUI Program Debugging Workflow Problem Solved When developing TUI (Terminal User Interface) programs, AI code changes require repeated manual testing. TUI-level feedback is particularly painful — either manual screenshots or typing prompts, resulting in low efficiency.

Using the tui-mcp MCP server:

launch - Start TUI program screenshot - Capture PNG snapshot snapshot - Capture text snapshot send_keys/send_text - Send keyboard/text input wait_for_text/wait_for_idle - Wait for pattern or terminal idle resize - Resize terminal to test responsive layouts

2. agc-explore — CodeGraph-First Exploration Problem Solved When subagents explore code, they often nest grep/read calls repeatedly, resulting in low efficiency. CodeGraph can return in one call what would take dozens of grep+read iterations.

3. agc-refactor — Complex Refactoring Workflow Problem Solved Complex design refactoring requires multi-branch reasoning, iterative verification, and visual snapshot comparison. Regular refactoring often leaves you wondering "did I get this right?"

Architecture CodeGraph → Yggdrasil MCP (multi-branch reasoning + session persistence) → TUIdbug (visual verification) ↓ Shell Scripts (state management) ↓ Git Worktree (on-demand isolation) Core Scripts Script Purpose check-mcp-config.sh Check and install required MCP configurations refactor-state.sh Manage refactor session state screenshot-manager.sh Screenshot capture, comparison, archival yggdrasil-helper.sh Yggdrasil session management Dependencies MCP servers: yggdrasil-mcp, tui-mcp, codegraph Sub-skills: using-git-worktrees, systematic-debugging 4. agc-docs — Project Documentation Standards Problem Solved Generic documentation standards don't match actual project needs. Every time you write docs, you have to纠结 format, naming, and content requirements.

Core Approach Instead of applying generic rules:

Interview user to understand project context Generate project-specific docs/DOC-SPEC.md Enforce project standards throughout development lifecycle Initialization Questionnaire Question Options Project type Application / Library&SDK / Infrastructure / Mixed Team size Solo / Small team (2-10) / Medium/Large team (10+) Doc types README, CHANGELOG, API, ADR, User Guide... Special requirements Naming conventions, formatting preferences, tool requirements... Trigger Rules The following operations automatically check DOC-SPEC:

Creating/modifying .md files in docs/ Creating/modifying project root .md files Version releases PR creation Template Library

# Install npm install @juntz/coding-skills

Cultural Influence of Music as Propaganda (2025)

Created a free browser-based lyric video maker

AI-Native Firms

Rep Cammack says China has deployed 'digital twins' of every lawmaker

Tweet claiming Claude Fable 5 has helped disprove the Jacobian conjecture

Show HN: Local-first CLI to make Obsidian vaults searchable for AI agents

Gmail AI Inbox

Coding Skills Development Report

Tyler Cowen: The Future Belongs to AI Maniacs

How you use AI AS HELPER?

LLM Wiki Maintenance: Drift, Contradictions and Review

Lake Natron

WhaleTrack I track the biggest Polymarket bettors and grade them by consistency

The Brick Game, Through the Ages

Apple Is the King of AI and Nobody Knows It

Autonomous AI Intrusions Are Here: Lessons from the Hugging Face Compromise

Reverse Compilation techniques – Cristina Cifuentes(1994) [pdf]

Neural Drive, a SuperTuxKart World Model that runs in the browser

How the Best Hackers Learn Their Craft(2018)

The Cost and Network Limits of Space-Based AI Compute

Don't Predict, Prioritize: Rethinking GPU Reliability Assessment

Getting nerd sniped by the h-index

Show HN: Visual DSA

Ask HN: Which blog have you recently found to be your favorite?

Before and After October 7, 2023

Show HN: R3 – Review locally and send feedback back to your coding agent

LSU physicists create first room-temperature quantum material

The Accidental Discovery of Phencyclidine [video]

Smart People in AI Published a Plan a They Expect to Be Ignored

Claude Fable produced a counterexample to the Jacobian Conjecture

Show HN: PostgreSQL course with database running in your browser

Comments