harlan-zw

Lighthouse checks one page at a time. Your SEO problems are usually on the pages you didn't check. Unlighthouse runs Google Lighthouse on your entire site with one command: npx unlighthouse --site <your-site> Free and open source, MIT. harlan-zw/unlighthouse t.co

I used Opus 5.5 to build a site where you can talk to brundlefly: my half human, half fly mascot I used Claude Code to build a home for Brundlefly's half-human, half-fly mascot. Wet walls, eggs, and a creature you can talk to. Opus 5.5 did such a good job; I initially was trying to with Codex but didn't get the results I wanted. Opus 5.5 was able to mostly one-shot any direction I wanted to take with it and did surprisingly well animating, especially given the "3d model" is really just a 2d image with depth. Brundlefly is a collection of Agent Skills. They focus on human-sounding text and keeping jargon out of users UI. The site uses Nuxt, Vue, and Three.js. Click him to start a conversation. His replies include cues for his mood and movement in the scene. Early conversation choices use precomputed replies. Free text uses live generation through Cloudflare and Claude. You can try it out yourself here brundlefly.dev

I built another humanize text skill: 🪰 Brundlefly There are already skills like Humanizer (blader/humanizer) and Stop Slop (hardikpandya/stop-slop). Humanize/deslop skills weren't doing it for me, so I built Brundlefly. I was still seeing jargon and architectural terms leak into user UI. Calling a payslip an “artifact”, for example. 🪰 Brundlefly Agent skills for writing that's part human. - /im-part-human: edits for the reader, preserving meaning and voice, checking claims, and respecting attention and accessibility. - /slop-naming-recovery: checks what users actually do, then proposes names that fit the product. harlan-zw/brundlefly

🪰 Meet Brundlefly Agent skills for writing that's part human. /im-part-human: Understand the reader. Show empathy. Be accessible. Verify truthfulness. Respect attention. /slop-naming-recovery: Stop slop naming leaking to your users harlan-zw/brundlefly

Nuxt SEO v6 ✨ Nuxt SEO v6 is here. It includes 2 new modules: AI-ready and Skew Protection. Plus 14 new features, 42 bug fixes, and faster, smaller performance. Get started: nuxtseo.com Let me know if you have any questions on it or feedback; happy to help :)

I built a CLI and Skill to help Claude Code refactor faster with fewer tokens An IDE already knows how to rename a symbol and update its references. I wanted Claude Code to use those operations through its shell. So I built RipIDE, a free, open-source CLI and Skill for symbol renames, moving exports, and updating imports. I used Claude Code to help replace ts-morph with TypeScript’s native language server. That meant connecting four tools: - ripgrep to find candidate files - TypeScript’s native server for references and rename edits - Oxc for parsing and syntax transforms - Volar for Vue scripts and templates One trap during that work: the server could rename a declaration but preserve its old re-export name. Consumers kept using the old name. The rename preferences had to go through workspace/configuration before requests, so they actually took effect. That migration and the configuration detail are recorded in this commit. Getting the tools to agree was harder than I expected. Commands preview changes by default. Supported refactors check for new type errors before applying. harlan-zw/ripide Any refactoring cases that you think would be useful :)

I built RipIDE to help agents refactor faster with fewer tokens I built RipIDE to help coding agents refactor faster with fewer tokens. Benchmarks (https://github.com/harlan-zw/ripide/blob/main/bench/README.md) indicate 45–60% fewer tokens and 35–60% less time across different refactoring work loads. One command handles references and edits across files, using TypeScript’s native language server, Oxc and ripgrep. It can rename symbols, move exports, update imports and rename CSS classes. Commands preview the diff by default. Supported refactors check for new type errors before applying. Free and open source, with an Agent Skill included. harlan-zw/ripide Let me know what you think :)

@SheltonLouisGT The broken templates are less of an issue these days; it's true, thank god. This is based on evals on fairly medium-to-large-sized projects running 6.1 Luna and GLM 5.3. I'll get some benchmarks with Opus 5.5 once my weekly renews. You can see here github.com/harlan-zw/ripast/tree/main/bench

Agents suck at refactoring. They're slow, break templates, miss declarations. Introducing 𝚛𝚒𝚙𝚊𝚜𝚝. A CLI with IDE-like refactoring. 🔸~50% fewer tokens, ~50% less time waiting 🔸TypeScript, React, Vue, Nuxt, Solid 🔸Components, CSS classes + vars harlan-zw/ripast t.co

☁️ Mdream v2 is out! HTML → Markdown in JS and Rust. Faster conversion. Smaller bundles. Less memory. Tree-shakable plugins. Stable as ever. Powers 🔸Cloudflare's /markdown endpoint and Markdown for Agents. harlan-zw/mdream t.co