Delegate bounded coding tasks to Antigravity CLI and review

Hands a bounded coding task to a separate implementer - the Google Antigravity CLI (agy) - then lets you review what it produced and land it yourself. You write the brief and own the judgment; Antigravity does the typing in its own conversation; you verify and commit. The loop needs only the ability to run a shell command and read a file, so any comparable agent can drive it - designed for and run on Claude Code, with other orchestrators treated as designed-for but not yet proven.

The loop is five steps. Steps 1, 4, and 5 are judgment calls; 2 and 3 are mechanical:

1. Write the brief. Antigravity sees only the text sent to it plus what it can inspect in the workspace - no chat history, no shared context. The brief needs the goal, the current state, what to change, what to leave untouched, the project's actual gate commands, and a report contract. Tell Antigravity it will not commit, and keep one task per brief.
2. Dispatch the brief with the bundled helper, which wraps agy --print, captures the run, and writes a structured result.json. The helper starts a fresh Antigravity project by default and passes --add-dir <repo> (the absolute --cd path) so agy has an explicit workspace, and it does not pass --dangerously-skip-permissions by default.
3. Wait for completion. The helper blocks until Antigravity finishes. A run is finished when result.json is written and the process has exited - trust that over any progress tracker. The implementer's full report is the finalMessage field in result.json, also printed in full on stdout between the report markers.
4. Review - do not trust the self-report. Re-run the project's gates yourself rather than accepting Antigravity's own gate claims. Read the diff against the brief - did Antigravity do what was asked, nothing more and nothing less - starting from touchedFiles. Run relevant guard skills on the diff if installed. For schema/migration changes, round-trip them; for removals, grep for dangling references.
5. Land it. The implementer edits the working tree; the orchestrator commits. Only after the gates pass and the diff holds: commit with a clear message, or send a delta brief with --resume-last for rework and review again.

Use it to hand off a bounded coding task to Antigravity and review the diff afterward.

Do NOT use it when the task is small enough to do inline, when the agy CLI is not installed or authenticated (install from Antigravity's CLI docs and complete first-launch setup), or when only Antigravity's opinion on existing code is needed - a --read-only dispatch covers review without edits, but a plain review may not need delegation at all.

Prerequisites: agy help succeeds (if not, install the Antigravity CLI and complete first-launch setup), agy models succeeds (proving the CLI can authenticate and list available model labels), and you are in, or point --cd at, the target git repository. These checks do not prove a headless write will be approved - in --print mode Antigravity cannot prompt for a write permission and may auto-deny it, which the relay detects instead of reporting completion.

agy has a configured default model, so --model is optional - use it when the human has a preferred Antigravity model label for the task, otherwise let Antigravity use its own current default rather than guessing.

node "<skill-dir>/scripts/relay.mjs" --brief brief.txt --cd /path/to/repo

Additional flags: --model "<label from agy models>" chooses a model label, --effort high sets reasoning effort (low, medium, high), --read-only runs plan mode with no edits, --sandbox enables Antigravity's terminal sandbox for the run, and --resume-last continues the most recent conversation with a delta brief.

Orchestrating agents or developers who want to delegate a bounded coding task to a separate Google Antigravity implementer session while keeping review and commit authority themselves.