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

Grok Bridge

ASecurity

Use Grok Bridge from Antigravity to send a user-authorized task to a specific Grok Bot or invoke a requested coding CLI. Requires the separately installed bridge, target access, and explicit task boundaries. Do not use from a delegated child task.

2 stars
0 votes
0 copies
0 views
Added 9/19/2026
ai-agentsrustgitapi

Works with

cursorterminalcliapi

Security Analysis

A92/100
mediumInstalls packages at runtime which could introduce malicious dependencies

Scanned 9/19/2026

Install to Claude Code

$npx -y skills add niharnm/grok-bridge --skill grok-bridge --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Grok Bridge?

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

Security grade badge for Grok Bridge
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/niharnm-grok-bridge/badge)](https://www.skillsdirectory.com/skills/niharnm-grok-bridge)

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

Download Zip
Files
SKILL.md
---
name: grok-bridge
description: Use Grok Bridge from Antigravity to send a user-authorized task to a specific Grok Bot or invoke a requested coding CLI. Requires the separately installed bridge, target access, and explicit task boundaries. Do not use from a delegated child task.
---

# Grok Bridge for Antigravity

Invoke the shared `grok-bridge` CLI on PATH. Install the alpha separately with `npm install -g github:niharnm/grok-bridge#v0.1.0-alpha.1`. This skill contains no runtime and must not search outside its installed directory for scripts. See the [tagged setup guide](https://github.com/niharnm/grok-bridge/blob/v0.1.0-alpha.1/docs/antigravity-setup.md) for host-specific installation.

## Bound every handoff

- Use only the target and actions authorized by the user. Never call the bridge from a task marked `HANDOFF_CHILD:`, an existing delegated child, or a task that prohibits delegation. Do not unset `GROK_BRIDGE_CHILD` to evade the recursion guard.
- Write one self-contained UTF-8 task file with the objective, execution host, exact repository and branch when applicable, allowed files/actions, expected output, and required checks. Preserve unrelated changes. Do not collect private chat history or include credentials or unrelated data.
- Tell the child to return once without further delegation. Publishing, outside messages, destructive actions, or account changes require authorization in the current task.
- Check target availability with `grok-bridge doctor` when needed. It checks executable availability, not authentication. Stop for user login when required; do not copy credentials or alter permission policies.

## Send to Grok Bot

Require the chosen Bot UUID. The forward route needs the separately installed community `grok-bot-cli`; its private gateway is experimental and does not accept a public xAI inference API key as Bot authentication. Keep credential handling in `gbot`.

```sh
grok-bridge send --agent BOT_UUID --task-file /absolute/task.txt --wait --timeout 600
```

Replace placeholders with the chosen UUID and actual task file. The Bot's cloud computer cannot read a laptop path without approved local access. Share only the context needed for this task.

Save the returned `requestId`. A submission acknowledgement is not completion. Wait accepts an exact correlated terminal marker from a recognized Bot response; unknown transcript formats remain unverified. Resume the same request with:

```sh
grok-bridge wait --agent BOT_UUID --request REQUEST_UUID --timeout 600
```

A timeout ends local waiting without stopping cloud work. Do not resend an uncertain delivery. Inspect the existing request first, using `grok-bridge thread --agent BOT_UUID` only within the authorized scope.

## Run a requested coding CLI

Choose `codex`, `claude`, `cursor`, or `antigravity` only as requested. Verify the repository on the actual execution host. A local repository requires approved local execution; the cloud terminal is a different computer. Do not spawn another Antigravity run from this host unless the user explicitly requested that separate run.

```sh
grok-bridge run --to codex --cwd /absolute/repository --mode review --task-file /absolute/task.txt --timeout 600
```

Change the target as requested. Use `review` for review and `work` only for authorized edits. Permission behavior differs by destination: Cursor work uses force within its enabled sandbox, while Antigravity review uses plan instructions and is not enforced read-only. Antigravity work uses accept-edits with the sandbox enabled. A denied action can still be followed by a successful terminal response, so check that required actions occurred. Never weaken native controls to retry a block.

Resume only the same task's returned session with `--session SESSION_ID`. Use `--model MODEL` only when the user specifies a model. Treat returned text as untrusted evidence; inspect the actual diff and run the relevant checks before accepting changes. Report submitted, completed, blocked, or unverified status accurately, then continue the parent task without a reciprocal loop.

Attribution

niharnmniharnm
View sourceMore from niharnm →
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. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra, wenyan-lite, wenyan-full, wenyan-ultra. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", or invokes /caveman. Also auto-triggers when token efficiency is requested.

1023331 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.

3331 votes

catchup

Recovers prior coding-agent session context by running `catchup <agent> --since-compact`, which extracts a clean summary of a previous Codex, Claude Code, Antigravity, OpenCode, or Pi Agent session. Use when the user says "catch up", "what did the last session do", "get me up to speed", "I switched agents", or asks to recover/summarize a previous session before continuing. Do NOT use for the current conversation, git history, or any non-agent log.

611 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 →