Got Tired of AI agents cheating tests, ran a little experiment..
I planted 8 cheats in small fixtures: tests skipped, commented out, deleted or rewritten to pass, assert(true), an untested branch. Plain CI was green on all 8. Aver passed none.
Aver judges an agent's change by the original tests. MIT.
I tested it on 12 of my apps and 4 open source and a few test apps needs give it a go and see if it helps you be compliant with the App Store
https://t.co/DvQLRDmI6I
Even If you pass the app review you can still get banned.
I have 12 apps on the AppStore, with developer accounts being banned left and right recently I’m scared I missed something somewhere.
I built a skill to review my apps to make sure they are compliant with this advice shared here by Phillip.
8 of my apps are AI based, apparently you need a report button on the output I wasn’t aware of which could get my account banned.
If you have an Apple developer account or an app in the App Store, read this now and share it.
Apple rejects your app. The notice is vague, so you fix what you reasonably think is wrong and resubmit.
Apple approves it.
Months later, with nothing in review, Apple terminates your account under 3.2(f) for “a pattern of manipulative or misleading behavior.” Every app gone. No explanation of what you did.
Your crime? Trusting the approval.
Apple’s own reviewer looked at your fix, with your rejection history in front of them, and said yes. Then Apple turned that yes into evidence against you.
So who’s misleading whom?
I ran App Review for seven years and wrote the guidelines. 3.2(f) used to mean real fraud. Now it’s being used against developers who did exactly what Apple approved.
I’ve made this Substack post free for everyone. Every developer needs to see it before this happens to them.
Share it before someone you know gets that email.
https://t.co/mWa1oGRVFH
2 hours in i'm now at 59%
19% usage used.
Model: Opus 5.5 High
currently:
- rewriting a old 90's RTS into rust
- agents powering though one of my apps roadmaps
- creating a new skill
mac cpu maxed
9 Hours left to max my ultra weekly usage can i do it?
Set Claude Code to high effort, and now every subagent burns tokens at high too?
Give them their own level.
Add this inside "env" in ~/.claude/settings.json:
"CLAUDE_CODE_SUBAGENT_EFFORT": "low"
Pasted a huge error log into Claude Code earlier and want that prompt back?
Press Ctrl+R and type a word from the log.
History search now matches text inside collapsed [Pasted text] blocks too. (v2.1.314+)
Forgot to book GTA 6 release day off?
No problem, Send your laptop in instead.
/gta-6 is a free Claude Code skill:
→ fake deploys scroll all day
→ mouse moves 1px every 4 min, Slack stays green
→ goes "away" at lunch, logs off at 5:30
→ writes your standup
/gta-6 stop → MISSION PASSED
@mattpocockuk A big problem I am seeing in the industry right now is, more PRs, more code, less time for testing and review, makes the higher ups happy with the metrics but now I am more incidents happening as result do this.
Does your course address this?
Neuralink’s first product is called Telepathy.
It enables high-speed communication with computers or humans simply by thinking. Output bandwidth will vastly exceed the fastest human speech or typing in a few years. Read that sentence again.
Started Claude Code in the wrong folder, deep into a conversation?
Don't quit. Type:
/cd ~/path/to/the/right/repo
The session moves there with the whole conversation, and picks up that folder's CLAUDE.md, settings and MCP servers. (v2.1.246+)