
I started developing UpdatesBee back in 2020, only it wasn’t called that then. Before that, I sold my stake in Easy Updates Manager to the kind folks behind Team Updraft when it hit 100,000 installs, helped develop its premium features with another developer, and later released a premium version. I was thankful for the contract at the time, and still am.
All was quiet until WordPress 5.5, when automatic updates for plugins and themes were going to be made native to Core. With extensive knowledge of the updates system, I volunteered to help develop and test the update mechanism in WordPress Core, and Updraft called on me again to make sure EUM was ready for the Core release. My recommendation was short. I believed there was nothing wrong with EUM’s approach and no need to change anything, which surprised David, the lead developer behind the popular backup solution UpdraftPlus. I believed EUM was still doing things the right way, even if updates were being brought natively into Core. David read my recommendations and ultimately agreed.
After WordPress 5.5 was released, things went quiet again, and I continued using EUM Premium on most of my sites with no ill effects. In a way, I went from contractor to customer, purchasing the software I helped develop because of a non-compete and Updraft’s policy of not benefiting privately from their hard-developed software (fair enough).
By 2020, my non-compete had long expired. I was paying the yearly subscription for EUM Premium and had ideas on how to improve it, but Updraft wasn’t interested due to budget constraints. So I decided to write my own version called Ultimate Auto Updates.
As you can see from the screenshot, it was almost a mirror image of EUM, but with extra features that made it a viable competitor (or at least, I thought)
But then imposter syndrome crept in, and I didn’t think I could develop a solution on my own to justify the “Ultimate” name. I wanted to code the interface in pure React, but I didn’t have the skills or time to develop the product beyond a few finished screens. That, and it being so close to EUM’s feature set, I decided to move on to other projects.
I would sometimes open up the Ultimate Auto Updates project, get overwhelmed by how much was left to develop it, and put it back on the shelf again. This happened repeatedly.
Fast forward to 2026. By this time, I was a seasoned React developer, having coded most of my plugins in the framework and learning through trial and error what worked and what didn’t. AI was prevalent and made developing React apps and UIs much, much easier.
So I dusted Ultimate Auto Updates off again in September of 2026, but this time I didn’t get overwhelmed. I was motivated. I hated the UAU UI, so I began sketching wireframes for what a viable fork of EUM might be, researched new names for the product, and considered what features it would take to release something I would want to use instead.
I settled on the name UpdateBee, which I thought described the product well. I purchased the domain and brainstormed on a mind map what was left, what features I wanted to bring to the table, and which ones I would fork.
I landed on the idea of a hard fork of EUM and another popular plugin called Companion Auto Update. I had long been a fan of CAU’s cleaner design and architecture. With my mind map of features, I began designing a new UI matching the UpdateBee name. I dropped the top navigation in favor of a sidebar. I began porting features from all three plugins, using a combination of traditional development, WordPress’ new (and still experimental) @wordpress/ui component library, and AI.
Developing UpdateBee grew like wildfire within me, especially because I had joined the unemployment line again after being laid off twice in a row because of AI. That’s when I decided to develop UpdateBee under the guise of releasing it for free on the WordPress plugin directory.
I asked my close friend and life mate Pauli if it was worth a public release, knowing there are many other update plugins available. She told me that if I have something unique to share, I should share it. In other words, to hell with my imposter syndrome, and to suck it up and go for it.
Knowing .org was in my future, I added its AI skills to my public repo and made a serious attempt to meet .org’s ever-stricter guidelines. I had just released another plugin on .org, and it was wild how much stricter it had become since my last free release. I thought, fair enough. Because of AI, more plugins than ever are being submitted, and because of AI, an ever-growing pool of security updates is being discovered (understatement of the year, I know).
As I wrapped up my MVP, I began cleanup. My motto is make it work, then make it pretty. This included hardening (or my attempts at it), endless bug fixing, and testing and testing some more (often on my live sites because that’s how I roll and find the real bugs). I finished up my feature set, fleshed out a comprehensive readme.txt and began what I imagined would be a tough .org review. I made sure PHPStan and Plugin Check (PCP) passed with flying colors first, and submitted it on Halloween (October 31st).
I was disappointed almost immediately. My submission attempt was hit with an automatic rejection. I couldn’t even upload it for review. According to the automated check, it violated Section 8 of the detailed plugin guidelines.
Section 8 reads: Plugins may not send executable code via third-party systems.
It goes on to say:
I read it as both an alarm bell and a false alarm. I wasn’t sure why it mentioned updaters, since there’s no clear .org guideline against them.
Not being able to submit dismayed me, so I emailed the plugins team at their catch-all group address (plugins at wordpress.org).
Hi plugin team,
Are update manager plugins no longer allowed on .org?
I’m getting this error on upload and I believe it to be a false positive as I’m not using any checks that phone home or alter the update mechanism significantly.
I then linked to my public GitHub repo, where I shared the latest version of the public code and ZIP file.
The response I received was encouraging and was from a human (name redacted for privacy) on November 5th:
Update-management plugins are not categorically prohibited. However, plugins may not download, install, or update plugins, themes, or other executable code from sources outside WordPress.org.
I mentioned I couldn’t get past the error, and I was gratefully emailed back a special link that let me bypass it and finally submit my plugin (same day, human service – impressive!). I began the wait, as there were hundreds of plugins in the review queue at the time.
The first response I received from a reviewer came on November 10th, and it was an encouraging sign (only five days after my eventual successful submission). The review gave me the impression that my slug had been approved (I was elated not to play the name game again). I took it in good faith, as my last submitted plugin had to be renamed several times before the review would even begin.
This mostly automated (and heavily AI) review flagged some Abilities API items that required removal, some prefixing errors, links to domains that didn’t exist in the readme, and loading Core files directly (like the versions.php file in wp-includes. So I stripped those out, fixed the errors/issues, and resubmitted 3 days later, on November 13th. I also addressed several ABSPATH checks I had left out, but that the reviewer didn’t flag, which I found odd. It took me three days to make the recommended changes, even with AI help, but I was confident the next review would be gentler.
Since I was waiting again, I used this time to start populating my domain and documentation, knowing my name and slug would stick, and preparing for the eventual public release. Although confident, I knew the review would take a while for a plugin of its size and number of moving parts. I moved on to other plugins and checked my email periodically, hoping for a response to my latest upload.
The next round of review came only five days later, on November 18th. I knew I was in for a ride, but the opening of the email was encouraging: “It’s time to move forward with the plugin review “ronalfy”!”
The review was getting stricter, but that comes with the territory. I was using constants to set some versions. Fair enough. Another line flagged a misuse of wp_enqueue_script. Basic stuff.
Then the red flag came.
Hidden amongst a sea of good advice and checks was a sentence of doom.
My heart sank. If you know anything about “update plugins”, it’s basically required to have these in the plugin. The review encouraged me to address the issues and resubmit.
They were asking me to remove the heart and the plugin’s main purpose. I appealed immediately.
Then on November 20th, the nail in the coffin was delivered:
The line that doomed me was:
My initial response upon reading that was a memory of the scene in We Were Soldiers, when the helicopter pilot yelled, “No shit,” as he took enemy fire trying to land at a hot landing zone. “No Shit,” exactly.
In that moment, I knew my review was doomed. There was no appealing that.
I thanked the reviewer (or AI overlord) for their time and asked that the review be discontinued.
On November 20th, I received a response. It began with another promising start: “👋 Hi and welcome to your plugin review ‘ ronalfy ‘!”
It then repeated everything in the last email verbatim. My heart sank even further, knowing it probably wasn’t even a human who sent it.
I confided in my long-time friend about how disappointed I was with the review, and how I longed for famed plugin reviewer Mika, and how I felt like I had hit an AI brick wall on .org.
His short, blunt response encouraged me: “Then sell it.”
In that moment, I realized I would take a long break from releasing my next plugin, or at least wait until things improved regarding the review process. I was burnt out and saddened. I had been submitting plugins since 2007, and never felt so betrayed. I could accept the rejection, and I ultimately did. Still, I had experienced firsthand what I had long read about others having: a miserable .org experience submitting their plugin and getting rejected.
The time I spent on the plugin was a sunk cost. Had I researched and prodded much earlier, a lot of this could have admittedly been avoided, but as someone wise once told me, “It is what it is.”
That’s when my girlfriend joined my good friend in solidarity, albeit separately: “Then sell it.”
I wanted to release the plugin to the community, and perhaps save others some costs down the road, but alas, it was not to be. Without .org, distribution is a moat that I have historically not had a toehold in.
So I took my good friend’s (and my girlfriend’s) advice: I prepared the plugin for sale, while pledging to at least do the community a solid and keep the source open.
I wish there were a silver lining in all of this, but the future doesn’t write itself in the present. I just wanted to share that not all rejections are created equal. This one hurt, and it was likely AI again that showed me the door.
If you’ve made it this far, you deserve a drink (whiskey) for reading. I promise the next dev diary will end on a positive note.
Cheers.
Ronald Huereca











