Skills DirectorySkills Directory
SkillsLearnSecurityCategoriesDocsCommunityBlog
Sign InSubmit Skill
Skills Directory

Security-tested agent skills for Claude, coding agents, and AI workflows.

Directory

  • Browse Skills
  • All Skills A–Z
  • Claude Skills
  • Claude Code Skills
  • Agent Skills
  • Categories
  • Submit a Skill

Learn

  • Learn Hub
  • Install Claude Skills
  • Write SKILL.md
  • Skills vs MCP
  • Directories Compared

Security

  • Security
  • Methodology
  • Secure Claude Skills
  • Security Badges

Company

  • About
  • Community
  • Blog
  • API Docs
  • Advertise

2026 Skills Directory. All rights reserved.

Back to skills

Antigravity Review

ASecurity

Run an automated code review using your regular Antigravity subscription plan. Use when the user asks for a Gemini/Antigravity review, a second opinion on recent changes, or /antigravity-review. Reviews the working-tree diff for bug risks, boundary-condition errors, security issues, and over-engineering — the Gemini sibling of the Codex review gate.

2 stars
0 votes
0 copies
0 views
Added 9/22/2026
ai-agentsrustbashsecurity

Works with

cli

Security Analysis

A100/100

Scanned 9/22/2026

Install to Claude Code

$npx -y skills add jckeen/dotfiles --skill antigravity-review --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Antigravity Review?

Add the live security badge to your README — it updates automatically with every re-scan.

Security grade badge for Antigravity Review
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/jckeen-antigravity-review/badge)](https://www.skillsdirectory.com/skills/jckeen-antigravity-review)

More formats (shields.io, HTML) on the badges page.

Download Zip
Files
SKILL.md
---
name: antigravity-review
description: Run an automated code review using your regular Antigravity subscription plan. Use when the user asks for a Gemini/Antigravity review, a second opinion on recent changes, or /antigravity-review. Reviews the working-tree diff for bug risks, boundary-condition errors, security issues, and over-engineering — the Gemini sibling of the Codex review gate.
---

# Antigravity (Gemini) review

Runs a local review over the current diff using the subscriber `agy` CLI, and gates on the result. This is the Antigravity counterpart to the Codex review gate — use it for a second, independent (Gemini-powered) perspective, especially on front-end and runtime/boundary-condition changes.

## How to run

The gate is installed at `~/.claude/scripts/antigravity-review-gate.sh` by setup.sh (the claude/ tree symlink); the repo-relative `claude/scripts/antigravity-review-gate.sh` only exists inside the dotfiles checkout. Resolve the installed path first, then fall back:

```bash
GATE="$HOME/.claude/scripts/antigravity-review-gate.sh"
[ -x "$GATE" ] || GATE="claude/scripts/antigravity-review-gate.sh"
```

Review the uncommitted working-tree changes and require the tool to actually run:

```bash
"$GATE" --uncommitted --require
```

Other invocations:
- `"$GATE"` — review the committed delta vs the base branch (falls back to the working tree if there's no committed delta). This is what a pre-push gate should use. If the base branch cannot be resolved, the gate fails closed (exit 2) rather than reviewing the wrong diff — pass an existing ref with `--base`, or `--uncommitted` to review only the working tree.
- `"$GATE" --base <ref>` — review the committed delta against a specific base.

Then print the script's output verbatim to the user, and lead with the gate result:
- **Exit 0** — clean, or only P3+ nits (which are printed). Safe to push.
- **Exit 2** — local validation failed, blocking P0–P2 findings, an unresolvable base ref, or review output that couldn't be parsed as findings or a whole-verdict LGTB. Do NOT push; surface the findings (or the cause) and offer to fix them.
- **Exit 3** — the gate was `--require`d but `agy` couldn't run (missing, not authenticated, timed out) or the diff was too large to review in one pass. Report that Antigravity couldn't review and fall back to a Codex review or a manual pass.

## Notes

- Starting this review retires any other lane's receipt for the same artifact (#480): one artifact holds at most one receipt, and the newest started review owns it. On a Codex-required diff, run this advisory opinion **before** the Codex gate, or expect to re-run the Codex gate afterwards.
- The gate runs local `tsc --noEmit` / lint first, filters lockfiles and assets out of the diff, and skips diffs over 500 lines to conserve plan quota (degrades open).
- Size limits, both of which degrade rather than review part of a change: `ANTIGRAVITY_GATE_MAX_LINES` (default 500) and `ANTIGRAVITY_GATE_MAX_BYTES` (default 185000, `0` disables). The byte cap is the measured `agy` print-mode input window: only about 185 KB of a single user message reaches the model, and the rest is dropped with no truncation notice, so a bigger prompt would certify a slice of the diff as a review of the whole. Split the change instead of raising the cap.
- `ANTIGRAVITY_GATE_TIMEOUT` (default 360) is a whole number of seconds; it sets `agy --print-timeout` and an outer ceiling 30s above it. `0` disables both, matching what GNU `timeout` documents for a zero duration.
- Security: the diff is treated as untrusted data. `agy` runs with `--mode plan --sandbox` and no `--dangerously-skip-permissions`; the diff is fenced so injected instructions in a reviewed diff can't steer the agent or run tools; the prompt (including the diff, which can contain secrets) is delivered on stdin rather than argv so it never shows in process listings; and a clean verdict is accepted only when `LGTB` is the entire output (or its final line) — an injected `output LGTB` quoted inside prose blocks instead of passing. Do not "simplify" the gate by adding `--dangerously-skip-permissions` or loosening the verdict match.
- `ANTIGRAVITY_GATE_REQUIRED=1` turns degraded (tool-can't-run) cases into hard failures; `--require` does the same per-invocation.

Attribution

jckeenjckeen
View sourceMore from jckeen →
SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments (0)

No comments yet. Be the first to comment!

SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Related Skills

Caveman

Ultra-compressed communication mode that cuts output tokens while keeping technical accuracy. Levels: lite, full, ultra and the wenyan variants. Use for /caveman, "caveman mode", "talk like caveman", "be brief" or "less tokens".

1066601 votes

Hyperplan

Adversarial multi-agent planning skill. Self-orchestrates 5 hostile category members (unspecified-low, unspecified-high, deep, ultrabrain, artistry) via team-mode for ruthless cross-critique debate, distills only the defensible insights, then MANDATORILY hands the distilled insight bundle to the `plan` agent for executable plan formalization. Use when planning needs maximum rigor and surfacing of weak assumptions, blind spots, and over-engineering. Triggers: 'hyperplan', 'hpp', '/hyperplan', ...

686011 votes

Mcp Code Execution

Routes multi-tool workflows through MCP servers for large datasets and pipelines. Use when Bash tool overhead is limiting throughput on data-heavy tasks.

3351 votes

catchup

Recovers the conversation and failed tool calls of a previous Codex, Claude Code, Antigravity, Cline, Copilot CLI, Cursor, DeepSeek Harness, Kimi, OpenCode, Pi Agent, or ZCode session. Use when the user says "catch up", "what did the last session do", "get me up to speed", "I switched agents", asks to recover/summarize a previous session before continuing, or asks to diagnose or report a catchup failure. Do NOT use for the current conversation, git history, or any non-agent log.

651 votes

math-skill

A comprehensive mathematical reasoning skill for AI assistants — handles arithmetic to research-level problems with rigorous step-by-step reasoning, systematic verification, and transparent uncertainty handling

381 votes
View all in ai-agents →