agentleFS
legibetlegibet/agent-skills/skills/archive/antigravity-cli/SKILL.md

antigravity-cli

Skill0 starsChanged 27 days agoMITView on GitHub

Delegate work to Antigravity CLI (agy) as an independent agent, or continue an existing agy conversation.

5 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 1 is distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 5 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Start it as a background command.

    line 21
  2. Very specific:

    Read /tmp/agy-<run>.json. Check in this order:

    line 30
  3. Very specific:

    Keep conversation_id to continue the thread.

    line 37
  4. Very specific:

    Run agy --help for the full command surface and agy models for other models.

    line 55

“Common” is measured: kept by at least 6 in 10 of the 5 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Antigravity CLI

agy -p runs one headless turn. The working directory is the agent's workspace; agy has no flag to set it, so cd into it in the same command. The final JSON envelope goes to stdout and the turn log to stderr; send both to per-run files.

Run a task

cd <workspace> && agy -p "<request>" \
  --model gemini-3.8-flash --effort high \
  --output-format json \
  --print-timeout 30m \
  --dangerously-skip-permissions \
  </dev/null >/tmp/agy-<run>.json 2>/tmp/agy-<run>.err
  • Start it as a background command. A run takes minutes, longer than a foreground shell call waits. The run is done when the process exits.
  • <run> is a name unique to this run, so parallel runs keep separate files.
  • --model gemini-3.8-flash requires --effort (low, medium, or high).
  • --dangerously-skip-permissions lets the agent run shell, URL, and MCP tools. Without it headless mode denies them.
  • --add-dir <path> adds another folder to the workspace.
  • For a long or multi-line request, write it to a file and pass < <file> in place of both -p "<request>" and </dev/null. Inline quoting breaks on quotes and newlines.

Read the result

Read /tmp/agy-<run>.json. Check in this order:

  1. status is ERROR (exit code 1): the run never started. error names the cause, usually a bad flag.
  2. /tmp/agy-<run>.err contains print timeout: the turn hit --print-timeout. status is still SUCCESS, but response is partial or empty. Continue the conversation to let it finish.
  3. denied_actions is present: tool calls were blocked, and the work they needed is missing.
  4. Otherwise response is the reply. When the agent waited on a background command, response opens with its progress notes; the result comes last.

Keep conversation_id to continue the thread.

Continue a conversation

cd <workspace> && agy -p "<follow-up>" \
  --conversation <conversation_id> \
  --model gemini-3.8-flash --effort high \
  --output-format json \
  --print-timeout 30m \
  --dangerously-skip-permissions \
  </dev/null >/tmp/agy-<run>.json 2>/tmp/agy-<run>.err

An unknown ID starts a new, empty conversation and still reports SUCCESS. Confirm the returned conversation_id matches the one you passed.

-c in place of --conversation <id> continues the most recent conversation in the working directory.

Run agy --help for the full command surface and agy models for other models.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.