The Day the Upload to Agent Overlay Took Every Drop, and How I Got File Handling Back
A full-window Upload to Agent overlay intercepts every drop, so files never reach the integrated terminal or the explorer sidebar. Here are the two workarounds the community found, plus three ways to hand over a path without dragging.
When an agent ends without a reason, I check the page's character encoding first
A page handed to ReadURL can end an Antigravity agent with nothing but Agent execution terminated. Here is how I tell those pages apart before handing them over, and why I now fetch and re-encode first.
When You Cannot Decide Between Rules and Skills: Rethinking Where Instructions Live
Rules and Skills load at different moments in Antigravity. Rules go in full on every turn; Skills lead with a heading and load the rest on demand. After the same note drifted apart in two places on a client site, here are the questions I use to place an instruction, plus a small script to measure what you carry each turn.
When the iOS 27 Date Landed, the First File I Grepped Was AGENTS.md
iOS 27 and iPadOS 27 ship on September 14. Old version numbers linger in your source, but also in the files only your agent reads. Here is a script that counts both zones, plus the one line keyword search will never find.
The File Counts Matched. 148 Pairs Disagreed on Code Blocks
My Japanese and English article trees matched at 1,057 files each, yet the bodies had quietly diverged. Here is what counting code blocks and H2 headings across every pair turned up, why a character-length ratio failed to catch any of it, and the 40-line checker I now run before every push.
When an Antigravity Scheduled Task Never Runs: Check sidecar.json, Then enabled, Then the Logs
A checking order for Antigravity 2.0 Scheduled Tasks that quietly do nothing: where sidecar.json lives, how the directory name becomes the ID, the enabled flag in config.json, and the logs under sidecar_data.
Before I accept what /boost says it verified, I count what its throwaway worktree never saw
Antigravity's /boost runs its tests inside an ephemeral isolated worktree. If your build reaches for files git never tracked, that verification passed on a different tree than yours. Here is how I measure the gap first.
When the few lines you quoted stop making sense in the next prompt
Antigravity 2.12.0 lets you quote part of a response into your next prompt, but the definitions those lines depend on stay behind. Here is a small script that counts what got left out, plus the widths I actually measured.
The Morning My Completions Stopped: Code Was 1.4% of the Repo
When completions stop coming back in Antigravity, restarting the language server often does nothing. Here is how I split the slowness into three layers, counted each one, and asked git whether my ignore patterns were actually matching.
With scheduled agents, I now look for runs that never happened before I look for failures
My run ledger showed a 100% success rate while the evening slot had not fired for two weeks. Here is the reconciliation I now run against an expected-fire table, with measured notes on cron expansion, exit codes, and how the ledger itself gets written.
A Custom Tool That Says any_of Loses Its Constraint Without Ever Erroring
I tested ten spelling variants in custom tool parameter definitions to see which ones fail loudly and which slip through unnoticed, then measured how argument validation shifts once normalization lands.
Pointing GOOGLE_GEMINI_BASE_URL at Your Own Gateway Does Not Change Who Holds the Key
I pointed the Antigravity CLI at a gateway of my own and recorded exactly which credentials travelled with it. Authorization is dropped across a cross-origin redirect, but the API key header survives every hop. Here is the measurement, and how I moved the key off my laptop.