A tools-and-workflow batch, practical pieces that save real time when applied.
madewithlove’s Running Multiple Claude Accounts Without Logging Out solves a real daily annoyance for anyone who uses Claude across personal and work accounts. The post covers browser profile separation, Claude Code’s config-based account switching, and the specific files and directories that hold authentication state. It’s short, practical, and immediately useful if you’ve been logging in and out repeatedly, which, judging by how many people I’ve seen complain about this, is most Claude Code power users.
Steve King’s Reviewing AI-Generated Work is a framework I’ve been wanting someone to formalize. The core insight: reviewing AI-generated code requires a different kind of attention than reviewing human code. With human code, you check for logic errors and style violations. With AI code, you’re also checking for plausible-but-wrong patterns, unnecessary abstractions the agent introduced to “be helpful,” and subtle misunderstandings of your codebase conventions that look correct in isolation but don’t fit the larger architecture. Steve proposes specific checkpoints and makes the counterintuitive argument that review time per PR should increase, not decrease, when agents write the code. I found this useful not just for my own reviews but as a reference to share with team members who are reviewing agent output for the first time.
Bots and Boundaries (Part 3) is the final installment of madewithlove’s series on handling automated traffic in production. This part focuses on the policy layer rather than the technical implementation, how to define, communicate, and enforce boundaries for bot traffic without accidentally blocking legitimate users or search engine crawlers. The technical details are Laravel-specific (middleware patterns, response code selection, escalation paths), but the policy thinking, what signals indicate abuse vs. legitimate automation, how to communicate limits without a wall of legalese, when to block vs. throttle, applies to any web application.
Excalidraw Frame Export describes a workflow for exporting specific frames from Excalidraw drawings rather than the entire canvas. If you use Excalidraw for diagrams in documentation or blog posts, this saves you from the screenshot-and-crop cycle every time you update a drawing. The post covers the Excalidraw frame API, automation options for CI pipelines, and integration with static site generators. Small but genuinely useful if Excalidraw is part of your documentation workflow.
Sebastian Bergmann’s From Events to Insights is a presentation about using PHPUnit’s event system to extract more than pass/fail from your test suite. By subscribing to test lifecycle events, test started, test finished, assertion made, error thrown, you can aggregate metrics across runs, track performance over time, detect flaky tests statistically, and build custom reporting that tells you what changed in your suite’s behavior, not just whether it’s green. If you maintain a large PHPUnit test suite and your current visibility into it is limited to “all green” or “something red,” this talk opens up a much richer observability layer.