Building open-source tools for the Claude Code ecosystem | Vet turned developer | Bangkok
I build open-source tools that make Claude Code better for everyone.
| Project | What it does |
|---|---|
| claude-code-statusline | Real-time rate limit %, context window %, session cost, and git status in your statusline |
| claude-code-notify | Push notifications to Mac, iPhone, and Apple Watch when Claude Code needs input |
| drawio-mcp | MCP server for creating draw.io diagrams from LLMs — XML, Mermaid, and CSV |
| awesome-claude-skills | Curated list of 100+ Claude Code skills across dev, data, DevOps, and more |
Reusable patterns and templates extracted from production AI systems I've built.
| Project | What it does |
|---|---|
| soul-agent-framework | Configure AI agents through markdown, not code — SOUL/MEMORY/skills architecture |
| ai-expert-chatbot | Deploy a domain-expert AI chatbot in minutes — RAG + SSE streaming + webchat |
| agent-factory | Spin up fully-deployed AI agents for any domain in under 2 minutes |
| openclaw-actual-budget | AI agent template — receipt scanning + expense tracking via Telegram + Actual Budget |
| telegram-order-dashboard | Telegram bot + zero-dependency Python dashboard for small business order management |
I'm a licensed veterinarian (Chulalongkorn University DVM) building AI tools for animal healthcare — a domain with almost no existing AI tooling.
| Project | What it does |
|---|---|
| vetblood-ai | AI-powered animal blood test analyzer for vet clinics — Gemini 3 Hackathon |
| tenacitOS | Mission control dashboard for monitoring and managing AI agent fleet |
GitHub | Kaggle Competition | Rank: #53 (private leaderboard, jumped +16 spots) | Score: 0.1587
Extracted structured voting data from 846 scanned Thai election result documents (Form สส.6/1) from the 2026 general election. Built a multi-model OCR pipeline with AI-powered assembly logic.
How we built it:
- Pipeline: 846 PNG scans → multi-model OCR → fuzzy party-name assembly → structured CSV
- OCR stack: Gemini 3 Flash + Gemini 2.5 Pro + Gemini 2.5 Flash, merged with official กกต. election data
- Key technique: Template-based party-name alignment instead of ballot-number mapping — solved the offset bug where 22+ docs had shifted ballot numbers
- Evaluation: Built a local TDD eval system (8 tests), but learned that 5-doc local eval had zero correlation with Kaggle — only empirical Kaggle testing worked
- Iterations: 16 Kaggle submissions, 6+ submission strategies, 900 OCR API calls (3-pass overnight run)
- Score journey: 0.929 → 0.307 → 0.228 → 0.158 (each jump from fixing a different root cause)
- Tools: Claude Code with
/loopfor 8+ hours of autonomous overnight monitoring
Used MiroFish (a multi-agent swarm intelligence engine) to simulate whether "vibe coding" would create a developer demand boom or bust over 2026-2028.
Simulation specs:
- 168 rounds, 40 AI agents, 1,484 total actions
- Agents represented startup founders, security engineers, rescue engineers, AI tool companies, regulators, and Thai SMEs
- Each agent had independent personality, long-term memory (Zep), and behavioral logic
- Seed data included 50+ real research citations (NYU, Stanford, Veracode, IBM breach reports)
Key finding: Developer demand bifurcates — senior security/rescue engineers see surging demand while junior prompt-only developers face commoditization. Not a uniform boom or bust, but a split.
- Doctor of Veterinary Medicine — Chulalongkorn University, former veterinary neurosurgeon
- Self-taught developer — shipped 30+ projects with TypeScript, Python, Next.js, and Claude
- Daily Claude Code user — building tools to improve the developer experience for the community
