The reading list
Essays, tech notes, and the in-between.
- technical-notes·3 min read
Raw SQL migrations: when they're enough, and the four cracks that force Alembic
Clipdex ships migrations as numbered `.sql` files applied by a Taskfile loop. That's been the right call for a teaching artifact. Here are the four specific cracks that will eventually force a move to Alembic — and the test for when 'eventually' becomes 'now'.
- AI·3 min read
Skills: The User-Facing Workflow Layer
Hooks enforce rules. Subagents review work. Commands and skills are how the team actually invokes them — thin wrappers route to detailed workflows, encoding multi-step knowledge in a single, memorable slash command. The architecture dissected.
- AI·3 min read
Subagents That Catch What Hooks Can't
Hooks block known-bad patterns. Subagents make the judgement calls hooks can't — type safety, architecture compliance, accessibility. One real subagent, dissected, plus a template for writing your own.
- AI·3 min read
The Identifier That Made My Podcast Sound Russian
I spent an evening shrinking ElevenLabs chunk size from 2500 characters down to 1000, convinced the model was choking on density. It wasn't. One camelCase identifier — appearing three times in a single paragraph — was making my voice clone drift into a Russian accent. A field note on debugging non-deterministic systems.
- AI·3 min read
From Audit to Hook: Turning Drift Into Enforcement
A tooling audit surfaces drift. CLAUDE.md tells the agent not to do it. Hooks make sure it actually doesn't. Eight real scripts from a production codebase, with the audit finding each one came from.
- AI·3 min read
The Claude Hooks Lifecycle: A Primer You Can Bookmark
Eight events, one JSON payload format, three exit codes that matter. The reference doc I wish I had open the first time I wrote a Claude Code hook.
- AI·3 min read
Tracking Tokens-Per-Session: The Number Claude Code Hides
Tokens-per-session is the single most useful signal for agentic coding sessions — and Claude Code doesn't show it. /usage gives you cost. /context gives you window percent. Neither sums to the number you actually want. Here's what to track, why, and how to put it on your statusline.
- AI·3 min read
Your Statusline is the Cheapest Feedback Loop in Agentic Coding
Most teams using Claude Code, Cursor, or Aider leave the statusline at default. It's a one-line strip of pixels that costs you nothing, sees zero tokens, and can change every decision you make during a session.
- technical-notes·3 min read
SMS Is a Lossy Channel: An Architect's Guide to OTPs, Alerts, and Graceful Degradation
SMS looks like a reliable transport because it usually works. Architecturally, it's a best-effort store-and-forward network with no end-to-end SLA. Here's what every engineer should design for before wiring OTPs, alerts, or 2FA to a text message.
- AI·3 min read
design.md, DESIGN.md, and Google Stitch: One File, Narrower Views
There's a popular framing that AGENTS.md is for coding agents and DESIGN.md is for design agents. That framing is a category error. Write one file; serve both.
