leek.io leek.io
contact
~/articles · 14 posts

Notes from the workbench.

Thinking out loud about PHP, AI tooling, ABA software, indie hacking, and whatever else I'm shipping this month.

I Went on Vacation and My Agents Paid Down My Tech Debt
2026·07·13 · ~7 min

I Went on Vacation and My Agents Paid Down My Tech Debt

I took a week off and let Superconductor loose on a 500k-line Laravel codebase. It opened roughly two dozen PRs a day: N+1 fixes, unbounded job processing, missing HTTP timeouts, tenant-scoping bugs. I merged 98% of them. This is the setup that made that safe.

#ai #tooling #methodology
Turning a Creator's X Tips Into Agent Rules Claude Actually Reads
2026·07·01 · ~8 min

Turning a Creator's X Tips Into Agent Rules Claude Actually Reads

I follow Povilas Korop for Laravel tips, but they scroll by on X and vanish. So I scraped ~696 of them, OCR'd the code-in-image samples with Claude, and distilled the durable ones into versioned agent-rules repos my coding agents load automatically. Here's the whole pipeline.

#ai #tooling #claude-code #tutorial
My Global CLAUDE.md: The Rules I've Built Up Over Time
2026·06·17 · ~11 min

My Global CLAUDE.md: The Rules I've Built Up Over Time

Every time an AI agent burned me, I wrote a rule. Here's my entire ~/.claude/CLAUDE.md — the personal rules that travel with me across every project — and the scar behind each one.

#ai #claude-code #tooling
A Linear workspace per project in Claude Code (no more reconnecting)
2026·06·06 · ~8 min

A Linear workspace per project in Claude Code (no more reconnecting)

If you work across two projects with two different Linear workspaces, Claude Code's Linear MCP makes you disconnect and reconnect every time you switch. Here's how to bind each repo to its own Linear account so it just works.

#ai #claude-code #mcp #tooling
Stop stuffing every AI rule into one root file: apply-agent-rules
2026·05·14 · ~6 min

Stop stuffing every AI rule into one root file: apply-agent-rules

Most projects pile every AI agent rule into a single root CLAUDE.md and the file balloons past a thousand lines. Subdirectory rule files are a much better idea — and apply-agent-rules makes them practical.

#ai #tooling
One LLM Isn't Enough: The Data
2026·04·24 · ~9 min

One LLM Isn't Enough: The Data

Two months of multi-model code review across 217 commits and 1,514 findings. Consensus is rarer than I thought, and 87% of real bugs were caught by exactly one model.

#ai #review #data
One LLM Isn't Enough
2026·04·03 · ~5 min

One LLM Isn't Enough

Relying on a single AI model to write and review your code is asking the student to grade their own homework. Here's the multi-model pipeline I run on every commit.

#ai #review
Getting More Out of Claude Code: Tips for New Users
2026·02·20 · ~3 min

Getting More Out of Claude Code: Tips for New Users

Treat Claude Code like a very fast junior developer who needs clear direction. A handful of habits — plan mode, scoped @ references, skills — make the difference between magic and mess.

#ai #claude-code
The New Skill in Software Development isn't Coding
2026·01·19 · ~5 min

The New Skill in Software Development isn't Coding

I shipped a web app to both app stores and ported a React app to native Swift for Apple TV — with no prior mobile, React, or Swift experience. The bottleneck isn't prompting. It's knowing what you want.

#ai #product #opinion
From Web to Living Room: Porting to Apple TV
2026·01·16 · ~5 min

From Web to Living Room: Porting to Apple TV

How I ported a React application to a native Swift game and shipped to Apple TV — without knowing Swift, or React. AI did the syntax. I knew what I wanted.

#ai #swift #product