Author

Scott Spence

@scottspence.dev

🤝 Svelte Society London co-founder 🧡 Svelte ambassador 📔 I blog what I learn over on scottspence.com 📈 Product Engineer engineering products 🛠️ Built svortie.com | devhub.party | skykit.blue | sveltest.dev 📍 UK

Posts

251

Publications

Readers

Scott SpenceOct 10, 2026

Revisiting LSP in Claude Code

Back in March I wrote about enabling LSP in Claude Code. It turned into one of my most-read posts this year, which is a bit awkward, because a fair chunk of it is now out of date. I told people to set an undocumented flag, I framed LSP as...

claude-codelsp

Scott SpenceSep 26, 2026

What Happened to My Coding Agent Guardrails After I Left

In June I wrote about how I stop LLMs drifting in production codebases. That post covered the checks I put in the repo so coding agents couldn’t quietly erode the architecture. What it couldn’t cover was whether any of it would survive without...

coding-agentssveltekit

Scott SpenceSep 24, 2026

omnirecall - Search Your Claude Code, Codex and Pi Sessions in One Place

Claude Code keeps your sessions on disk, for 30 days by default. Each one is a JSONL file in , one folder per project: every prompt, every response and every tool call. When I found that out, it gave me an idea. If I want to search my...

claude-codepi

Scott SpenceSep 22, 2026

Copy Buttons and Line Numbers in mdsvex With twinkleplop

I’ve had syntax highlighting on this blog since I moved it over to mdsvex back in 2021. What I’ve never had are the controls: a button to copy the code, highlighted lines, or line numbers you can turn on and off. I’ve finally sorted all three. Very...

sveltekitmarkdown

Scott SpenceSep 17, 2026

Give Coding Agents Your Project Docs With node:sqlite and FTS5

On a reinsurance platform I built at an agency this year, the documentation got big. Requirements, specs, schema notes, meeting notes and client emails. It went from 65 documents in May to 291. Coding agents had two ways to deal with that. Read...

sqlitenode

Scott SpenceSep 15, 2026

I Built 21 MCP Tools. These Are the 2 I Still Use

Since January 2025 I’ve built 21 MCP servers and tools. Most of them were experiments, a way of getting familiar with the technology while it was new. There are two I still get use out of every day. This is how that happened. January 2025

mcpclaude-code

Scott SpenceSep 9, 2026

Moving my development setup from EndeavourOS to Ubuntu

I created a development setup from EndeavourOS to Ubuntu 26.04. It’s for work: they need 1Password’s security check scripts running, and the drive has to be encrypted. My machine runs EndeavourOS, so I needed an Ubuntu environment alongside...

ubuntulinux

Scott SpenceAug 29, 2026

What my prompts say about how I work with AI

September 2023 is when I started getting paid to work with AI. Features had to work, bugs had to be fixed and pull requests had to survive review, in a large private client codebase where a plausible answer wasn’t the same as a correct one. The way...

claude-codepi

Scott SpenceAug 25, 2026

Agentic engineering: a practical guide to reliable coding agents

Agentic engineering is the practice of putting coding agents inside an engineering system that makes context, scope, validation, evidence and human review explicit. The model can plan and write the code, but it does not get to decide that its own...

aicoding-agents

Scott SpenceJul 25, 2026

Coding agent harnesses with my-pi

I have a habit of turning repeated arguments with coding agents into software. If I tell an agent to stay inside five files often enough, eventually I stop asking nicely and build something that blocks the sixth. If I keep asking for validation...

pimy-pi

Scott SpenceJul 25, 2026

How I work with LLMs

In November 2025 I wrote the honest version of working with Claude Code. At the time, most of my process involved spotting when Claude was drifting, calling out the bullshit, checking that it understood the task, and restarting the session when the...

aillms

Scott SpenceJun 21, 2026

How I Stop LLMs Drifting In Production Codebases

The problem I care about with LLM coding is drift: one plausible shortcut gets copied, then future sessions start treating it as how the app works. I’ve used AI coding tools every day as a team lead on large private client codebases, across...

pimy-pi