Peter Knezek

167 posts

Peter Knezek

@peter_knezek

Frontend dev • Building tools that save devs hours • har-to-mocks, mocks-to-msw • Tips on modern development, AI dev & productivity • GitHub: peterknezek

Katılım Kasım 2015

43 Takip Edilen21 Takipçiler

Sabitlenmiş Tweet

Peter Knezek@peter_knezek·19 Mar

When comments dropped, I got worried about code quality slipping... So I’m now building an AI-powered review automation tool. Our setup: self-hosted Bitbucket (not GitHub Cloud), so no ready-made plugins exist. It has to be custom-crafted from scratch!

English

206

Peter Knezek@peter_knezek·19 Nis

Oops, I missed the button to open the editor earlier 😔 Now that I have an Agent open with the browser on one screen and the editor with the actual implementation on the second screen… It's actually a pretty good combo. 😅

English

Peter Knezek@peter_knezek·15 Nis

Dislike the new Cursor update. Can't find essentials like current branch (now nested) & no ESLint warnings on code tab. Migrated to other tools but still used it. Might switch back to Code or Antigravity to check code. Anyone else noticing? 👇 #Cursor #AIDev #WebDev

English

103

Peter Knezek@peter_knezek·11 Nis

@mattpocockuk That's a good idea. Some companies are restricted from using Docker Desktop due to licensing and opt for Podman instead. btw: The sandcastle interface is elegantly crafted. Well done.

English

Matt Pocock@mattpocockuk·10 Nis

Thinking about moving Sandcastle off Docker and making the sandbox totally pluggable. I.e. Sandcastle would become an orchestrator that works with any coding agent in any sandbox - local or remote. Would love feedback here: github.com/mattpocock/san…

English

171

18.9K

Peter Knezek@peter_knezek·7 Nis

This is exactly where ESLint transforms into an active dev partner. It used to be a "set it and forget it" config file, but AI is finally unlocking its true potential. 🚀 #WebDev #ESLint #AI #SoftwareEngineering #CodingLife

Cory House@housecor

This is so slick. It used to be tricky to create an ESLint rule. Now I just tell Claude. Example: We want to assure all SQL queries are multi-line (to avoid merge conflicts). This rule reports anything that looks like a SQL query that isn't decorated with a SQL comment. The SQL comment triggers Prettier formatting via prettier-plugin-sql. Result: If I forget to decorate a SQL statement, linting autofixes it and Prettier auto-formats. Slick.

English

Peter Knezek@peter_knezek·2 Nis

Now the new standard is fully adopted. PRs are cleaner, and real benefits are showing. The “easy” quality improvements are never just one step. Anyone rolled out type checking / linting this way? How did it go? 👇 3/3 #CodeQuality #DevProcess #SoftwareEngineering

English

Peter Knezek@peter_knezek·2 Nis

Started by running the type check locally. Had to: • Fix all existing issues in the codebase • Prepare clean changes • Create a PR After the merge, there was a short “everything works” window… then new PRs started failing. 2/3

English

Peter Knezek@peter_knezek·2 Nis

Plan was simple: improve code quality by adding a missing type-check step. Thought it would be an easy task… turned out to be a full multi-step process 😅 1/3

English

Peter Knezek@peter_knezek·31 Mar

This is how I’m bringing real AI into locked-down enterprise setups. If your team runs self-hosted Bitbucket and hates losing review context, try it. #MCP #Bitbucket #EnterpriseAI #CodeReview

English

Peter Knezek@peter_knezek·31 Mar

Workflow in action 👇 1. Switch to your local branch 2. Ask your AI 3. Get a clean summary of every pending comment with direct file links This solves the exact “hunting comments” frustration I tweeted about. Review suggestions coming next. 2/3 github.com/peterknezek/bi…

English

Peter Knezek@peter_knezek·31 Mar

Remember when I said I’m building a custom AI review tool for self-hosted Bitbucket because comments were getting lost? I just shipped the first piece: an MCP server that lets your AI instantly fetch ALL unresolved PR comments with clickable links. 1/3

Peter Knezek@peter_knezek

English

Peter Knezek@peter_knezek·28 Mar

I can't wait for Claude Code for Real Engineers (Cohort 3) starting Monday! 🚀 I've been using Claude daily, big Nx monorepo refactors… but this cohort looks like the next level for real production workflows. #ClaudeAI #AIForEngineers aihero.dev

English

Peter Knezek retweetledi

Noah Zweben@noahzweben·26 Mar

Thrilled to announce Claude Code auto-fix – in the cloud. Web/Mobile sessions can now automatically follow PRs - fixing CI failures and addressing comments so that your PR is always green. This happens remotely so you can fully walk away and come back to a ready-to-go PR.

English

336

504

6.1K

1.5M

Peter Knezek@peter_knezek·20 Mar

Key takeaway: Mastering pwd is like learning to check your location on a map before you start driving. It’s the foundation of confident CLI use. Which one do you use the most?👇 #DevTools #Productivity

English

Peter Knezek@peter_knezek·20 Mar

Quick friendly tips: • The terminal is case-sensitive (pwd ≠ PWD) • If you type it wrong, just use backspace — no stress • Use it every time you open a new terminal window It becomes automatic after just a few days.

English

Peter Knezek@peter_knezek·20 Mar

First essential command in the series: pwd (Print Working Directory). It answers the #1 question every time you open the terminal: “Where am I right now?” Let me walk you through it like a simple presentation 👇 #Frontend #CLI #WebDev

Peter Knezek@peter_knezek

Surprisingly, so much of our work happens in the terminal — from project setup to build debugging. Starting a mini-series: #CLI essentials for #Frontend devs. Quick Q: Which bash/zsh commands do you use daily & think every FE dev should know? Daily one tip 👇 #WebDev

English

Peter Knezek@peter_knezek·20 Mar

TypeScript excels at inference — stop fighting it! 🔥 Redundant types aren’t “more explicit” — just more maintenance, noise & drift. If the compiler knows the type, let it. Future you will thank you. 👀💡 Still over-typing in 2026? Worst example below 👇 #TypeScript

English

Peter Knezek@peter_knezek·19 Mar

@smith_john54162 Yes... I’ll try to keep posting here. btw, what's your suggestion about it?

English

superhuman@smith_john54162·19 Mar

@peter_knezek Yeah, that makes a lot of sense. Going through everything first and then filtering later feels much closer to how real code reviews work, especially on larger codebases. Are you thinking of grouping the suggestions by type?

English

Peter Knezek@peter_knezek·19 Mar

Update on my code review struggles from last month 👇 ESLint helps a lot, but reviews still ate too much time. We started distributing reviews to every team member. PR acceptance speed jumped massively… but comments almost disappeared. 😬 #CodeReview #DevWorkflow

Peter Knezek@peter_knezek

Some days review eats >4 hours — massive context switching, I still refuse to merge without deep human review of design decisions. AI makes generating PRs easier than ever… Anyone, what actually works for you (tools, habits, prompts…)? Share below 👇 #CodeReview #DevWorkflow

English

Keşfet

@mattpocockuk @smith_john54162 @elonmusk @BarackObama @taylorswift13 @cristiano @BillGates @NASA