All Articles
Why Antigravity Agents Build Your Xcode Project Fine One Day and Break It the Next on a Synced Folder
Put an Xcode project inside a Dropbox-synced folder, hand it to an Antigravity agent, and builds start passing and failing at random. Here is the real cause — sync daemon versus build — and how to fix it with xattr exclusion and a conflict-copy preflight.
Turning a throwaway prompt into a reusable sub-agent
When a one-off prompt to an Antigravity 2.0 dynamic sub-agent works beautifully, it usually vanishes into your chat history. Here is how to capture it as a reusable definition, with the actual file layout and the distillation steps.
Hand Over Generation and Shipping, but Never the Signing Key — Designing Key Custody and Handover for an AI-Driven Pipeline
Even in an era where AI Studio and Antigravity take over everything from generation to internal-test shipping, the app signing key is in a class of its own. Lose it, and that app can never be updated again. As a solo developer who has run several apps for years, here is how I design key custody — separating the upload key from the app signing key, storing it, and planning the handover for the worst case.
Green in the Embedded Emulator, Broken on the First Real Device — Putting a Parity Gate on AI-Generated Compose Apps
AI Studio now generates Kotlin/Compose apps from a prompt, runs them in an embedded emulator, and pushes them to a real device over USB — all from one screen. Yet a screen that passed in the emulator can break the first time it lands on a real phone. As a solo developer running several apps, here is how I put a gate that catches device parity issues before they ship.
Designing MCP Tool Output So It Doesn't Flood Your Agent's Context
When a custom MCP server returns large results in one shot, your Antigravity agent quietly degrades. Field projection, pagination, resource_link, and an output budget keep context from overflowing — shown with concrete TypeScript and measured numbers.
Splitting a Giant Antigravity Diff Into Meaningful Commits
Are you committing the agent's 400-line diffs as a single blob? Here is a practical workflow, with scripts, for re-splitting an unreviewable bulk commit into one concern per commit.
An Agent Granted 'Watch an Ad to Unlock a Wallpaper' Entirely Client-Side — Re-Verifying Reward Grants with AdMob SSV
I asked an Antigravity agent to wire up 'watch a rewarded ad to unlock a wallpaper,' and it returned an implementation that wrote the unlock flag client-side only. Here is why that is not enough, how I re-verified the reward grant with AdMob server-side verification (SSV), and how I stopped double grants too.
Before a Major Update Silently Breaks Your Overnight Automation — Designing a Staged-Adoption Canary Gate
After a major update dropped my unattended run success rate from about 98% to 63% overnight, I built a staged-adoption gate that freezes the working setup, verifies a new version against a golden output in an isolated profile, and only then adopts it. Here is the design with bash and Python.
When Xcode Cloud's Free 25 Hours Quietly Run Out at Month's End — Field Notes on Measuring and Defending Compute Time
Before Xcode Cloud's free compute allowance drains at month's end and your builds sit queued, measure where it goes. Practical field notes on pulling usage from the App Store Connect API, stopping wasteful builds early, and trimming test runs to keep CI inside budget.
When an AI Studio App Won't Install on My Phone Because the Signatures Don't Match
You generate a Kotlin/Compose app in AI Studio, push it to a real device over USB, and the install dies with INSTALL_FAILED_UPDATE_INCOMPATIBLE. Here is why a debug-signed build can't overwrite your Play release, and how to test it without losing your production app.
A Translated Line Had Quietly Reverted to English — Guarding String Resources an Agent's Refactor Touched
Let an agent tidy your values folder and translated strings can silently revert to the source text. Here is a design and implementation that treats the default locale as the source of truth, reads every other locale as a diff, and blocks only dropped keys, reverted translations, and broken format arguments at pre-commit.
The Day Rolled Over and Yesterday's Wallpaper Came Back — Designing Date Boundaries Instead of Trusting an Agent's Time Code
A record of redesigning a daily-rotating wallpaper feature that showed yesterday's pick after midnight for some users and switched twice in one night for travelers. I separate the device clock, time zone, and daylight saving time into three distinct failure sources, show how to define a stable day key, defend reward boundaries against a rewound clock, and test it all with an injectable clock — from a solo indie developer's point of view.