The Founder & CEO of eLightUp and the creator of the widely adopted WordPress plugins Meta Box (@wpmetabox) and Slim SEO (@wpslimseo).

Hanoi, Vietnam
Joined May 2008
Anh Tran retweeted
We give @automattic a lot of shit, but the company has contributed the most sponsorship money to the PHP Foundation since 2021. So there's that.
12
36
7
314
18,722
Anh Tran retweeted
The PHP project has released fixes for 2 critical vulnerabilities affecting servers that host PHP applications. Site owners and technical teams should confirm their hosting provider’s update plan and update PHP immediately. Versions before PHP 8.2 no longer receive security updates. These issues are not specific to WordPress, and no WordPress update is required. wp.me/pZhYe-5Gb
93
1,316
155
3,594
308,338
Moved our entire hubspot blog over to wordpress using claude code which ran autonomously for 6 hours and emailed me when it was done. As people become more savvy, the "platform migration PITA as customer lock" goes away. Here's how I did it. First I used AI to find the most robust MCP out there for a blog host. WordPress beat webflow for my use case. Now the fun begins, I woke up around 5:30am and on the walk portion of my morning run I kicked off deep research for documentation for both hubspot and WordPress, given I was going to migrate. Got home showered, watched the kiddos walk to the bus stop, then went to work, got in by 8:15am. Hubspot and WordPress MCPs were already connected (I used a business account for a new WordPress site to get the full mcp) While at my desk I told Claude code that I wasn't going to be at my desk after 9:30am and that it would have to work autonomously from 9:30am until I got back to my desk at 4pm. Knowing that for the next hour+ I had to get all the permissions set up. Step by step I installed the plugins, set up tilhe python, etc etc. I then left at 930 to go to my meeting in the city. I got an email at around 3pm in my drafts, telling me it was done. It ran for 6 hours autonomously, and when I show people the staging site, they can't tell which is which. The AI even caught that it should move the forms and tracking over as embeds. If you look at these the only giveaway was accessibility icon. Not bad Claude :) Also now that it is on WordPress with the MCP guess what? I can voice things like this into my phone and get Claude to find an image and insert it. Prompt: Take this post and insert an image of me in the first image position and push the other image of cloudflare down, run indepentdly.
49
5
10
98
36,297
You might not need MCP for create/edit/delete posts.
WordPress has had a REST API since 2016, and it's one of the reasons the platform is so ready for AI agents. Every post, page, image, and user on a WordPress site has its own endpoint, which returns clean JSON instead of HTML an agent has to dig through. An agent can also figure the API out on its own. Request /wp-json/ and you get a list of every route on the site, including ones plugins added, and an OPTIONS request to any endpoint tells you what arguments it accepts. WordPress had this nearly a decade before anyone was talking about tool definitions for LLMs, and now it's perfectly set up to help agents engage with your content. elementor.com/blog/wordpress…
1
3
333
To create my claude.md, I had it read: - my favorite programming books + blogs - the best oss projects - the code of the engineers I respect the most Claude (when given proper instructions, constraints, and context) now writes better code than me, almost always.
the coding agent haters are getting dunning-kruger'd so hard right now do you really think that your engineering ability is better than the models at this point? it is very very likely that fable is better than you in every branch of software it has been trained on all the code of the internet, it knows almost every trick in the book, all the design patterns of every domain, every single low-level detail, every bit-twiddling hack, it's read every single blog post from every cracked engineer that ever existed do you really think you can beat it? do you really think you can be more creative than it? unless you are 99th percentile expert in some niche domain of software, it simply is not possible. sorry.
105
207
37
4,425
405,038
Anh Tran retweeted
Claude writes about 90% of Claude Code's code. From June to September, merged PRs more than doubled while the share reverted fell by more than 40%. We wrote about how we changed code review to get there, and how to adopt it: claude.dev/blog/solving-code…
417
894
468
12,772
1,808,574
Anh Tran retweeted
If you're like me and need to understand the skills and plugins you feed your agent, then you'll appreciate this deep dive of @poteto's pstack I had Opus 5.5 create. I was hesitant at first, but I found that it's a masterclass in agentic engineering. claude.ai/artifact/JNBnuspjG…
51
132
15
2,004
193,565
Anh Tran retweeted
something that keeps bothering me is how some people see slop as something that is happening to them, like they're some kind of helpless bystander powerless to do anything about it excuse me but you are a software engineer, you literally learned how to make sand think and make computers do your bidding?! and now you can't do anything about it? please the 2 things you should know to cut through all the noise: 1. you need to give your agent a way to verify its own work (i just posted about this so won't rehash) 2. your codebase is memory the 2nd is not new either, we have always had lint rules, compilers with good diagnostics and strong type systems, LSPs, etc to help us write better code. there is no difference when agents write the code instead, you just need more of them, and stronger ones. forget rules and skills. you can add those later. yes, your codebase is a form of memory to agents! agents love extending what already exists. so every nit and every annoyance you've ever seen, every mistake and bad pattern you observe from agents, needs to turn into a constraint, or the agents will just keep repeating those same mistakes like a virus in your code. then you keep observing and repeating this process until your codebase is locked down enough that even agents with little context can one shot a high quality PR. an even better direction is to refactor/rewrite your code such that agent anti-patterns become categorically impossible. use type systems to your advantage, use libraries/frameworks that reduce the architectural search space, or heck, create your own framework like we did for grok bot and design high quality primitives and well thought out APIs
the slop conversation is so one sided, it's not just that you can produce more code you can also observe, debug, and refactor more code even if the code is slop you have such good tools to deal with it. even if you write code manually, it's still 100x easier to maintain it
236
340
103
4,626
328,149
Anh Tran retweeted
🎯Upgrade the way to work and connect with Custom Models with: 🔹 Custom Model field: Select model items directly in a field. 🔹 Model object type in MB Relationships: Connect a model to posrs. terms, users or another models. Learn more 👇 metabox.io/custom-model-fiel…
1
4
118
Anh Tran retweeted
looking at pstack by @poteto and it's a goldmine of prompt engineering even her example prompt is the best template i've seen of the four things you need to give your agent in a single prompt: * a goal * a checkable finish condition * the evidence you want to see * constraints I'm always surprised by how short pro's prompts are. It's less about length, but ensuring your hitting those four points. Here's the prompt: /poteto-mode fix the duplicate notification after retry. Reproduce it first. Done means retrying produces exactly one notification and normal delivery still works. Show the failing and passing evidence. Leave a PR for review. So simple define what you want, how to know when you're done, what you want to see, and what you shouldn't do I guarantee you this isn't obvious to most prompters here's the sauce github.com/paperclipai/potet…
45
51
7
1,041
68,608
In a world where AI can generate layouts, mockups, and copy in seconds, it’s tempting to let automation do the heavy lifting. But if you want the best outcomes for your clients, you need to keep learning how to build and design because human judgment, context, and craft still win projects and relationships. Remember that clients hire you for more than output. They hire your expertise, your problem-solving, and your taste. Those come from practice and study, not from a plugin. Knowing fundamentals (typography, accessibility, information architecture, performance) lets you evaluate and guide AI outputs instead of blindly accepting them. Creative confidence helps you translate business goals into experiences. Use AI as an assistant, not a replacement. Let it generate drafts and ideas, then apply your judgment to refine and polish. Treat AI outputs like rough sketches. Check for context, feasibility, brand voice, and edge cases that AI might miss. Clients feel the difference when work is guided by insight rather than convenience. AI can speed up parts of the process, but it can’t replace curiosity and our judgment. Keep learning how to build and design so you can use AI where it helps and step in where it can’t. Your clients will get better results, and you’ll stay indispensable. #wordpress #webdesign
1
1
1
4
381
Or just switch it it @WordPress.
Every vibe coder's SEO prompt: "Google isn't showing my vibe-coded website. - Render it server-side. - Generate a sitemap. - Submit it to Search Console. - Unblock Googlebot in robots.txt. - Delete the noindex tags. - Kill the redirect chains. - Fix the 404s. - Add canonical tags. - Write a meta description for every page. - One H1 per page. - Add FAQ schema. - Add breadcrumbs. - Link the orphan pages. - Alt text on every image. - Convert images to WebP. - Fix the layout shifts. - Load in under 2 seconds. - Remove the obvious AI slop. - Add a real author bio. - Get quality backlinks. Rank #1 by Friday. Make no mistakes." Funny thing is, this is actually the whole SEO checklist. Bookmark it. Paste it into Claude Code. Then do the last two yourself.
25
15
7
180
76,459
Anh Tran retweeted
EmDash 1.0 is a stable, open source CMS built for Astro, with agent-friendly workflows, secure sandboxed plugins, and a decentralized registry that keeps publishers in control. cfl.re/4iXWjKx #BirthdayWeek
41
115
39
1,139
124,151
All WordPress users need to update their WordPress installations immediately 💯 There is a 9.2 CVE that discloses a remote code execution vulnerability that affects versions all the way back to 2016 😬 Here's the official notice github.com/WordPress/wordpre…
33
212
33
863
145,333
Cloudflare ha liberado su propia skill para hacer auditorías de seguridad con IA. Es la que utilizan internamente en la empresa para detectar vulnerabilidades y ahora cualquiera puede utilizarla con su agente. Se llama Security Audit Skill y organiza cada auditoría en 6 fases: ✓ Reconocimiento de la arquitectura y superficies de ataque ✓ Búsqueda de vulnerabilidades por cobertura ✓ Validación independiente de cada hallazgo ✓ Generación de resultados estructurados ✓ Verificación final con agentes distintos ✓ Creación automática del informe de seguridad Lo más interesante es que el agente que valida una vulnerabilidad nunca es el mismo que la encontró, reduciendo bastante los falsos positivos. Es completamente gratuita, open-source y tiene licencia MIT. Repo 👇
37
299
15
2,645
350,764
Anh Tran retweeted
What if you could generate a site with Lovable, Base44, v0, Claude, ChatGPT, etc, and import into WordPress as a real block theme? Oh wait you can github.com/Automattic/static…
16
28
2
166
13,338
Anh Tran retweeted
Unsure how to get started with Jev? Install the skill: npx skills add typesafe-ai/skills --skill typesafe-ai Then prompt in your project: use /typesafe-ai to see how Jev can be used to replace slow, expensive LLM usage and find possible new features it would enable for users.
36
120
12
2,594
143,886