delegate-skills
A set of skills that hand a task to a separate CLI agent while you review the diff and land the commit yourself
- 2,597 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- Launches third-party CLI agents that edit the working tree and run commands
- Some CLIs do not enforce a read-only mode
Install
- Works in
Manual install
npx skills add amElnagdy/delegate-skillsInstalls the whole delegation skill package.
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Launches third-party CLI agents that edit the working tree and run commands
- Some CLIs do not enforce a read-only mode
This is third-party code. Review the repository files before installing.
What it does
A package of *-delegate skills, each bound to one CLI agent: Claude Code, Codex, Cursor, Aider, OpenCode, Cline, Copilot and others. The orchestrator writes a brief, the skill launches the chosen CLI in a separate session through relay.mjs, waits for a structured result.json and returns the diff with a report. A separate delegate-setup skill discovers installed CLIs and organizes them into named lanes such as feature, tests and ui after you approve. Committing always stays with the reviewer: the relay never commits. The scripts use Node built-ins only, with no dependencies and no network calls.
Who it is for. For developers who work in one orchestrator and hand tasks to several CLI agents.
Good fit when
- You want to hand a bounded task to a separate agent and review its diff before committing
- You want to run a queue of tasks through different CLIs while staying the reviewer
- You want to fix which agent handles features, which handles tests, which handles UI
Not a fit when
- The task is small enough to do inline in the current session
- The required CLI agent is not installed or not authenticated
Example request
Use codex-delegate to have Codex rewrite the billing module, then show me the diff to review and commitLimitations
Each *-delegate skill needs its own CLI installed and authenticated as you would at the terminal. Node 18 or newer and git are required. Some CLIs have no enforced read-only mode: for those the diff and the touched-files list are enough to review against, but not a guarantee of isolation. The full delegation loop is verified on Claude Code; other orchestrators are designed for but not all proven.
How to disable. Remove the *-delegate skill folders from the agent's skills directory, for example .claude/skills/ or ~/.claude/skills/.
FAQ
Who makes the commit?
Always you. The relay launches the CLI and returns the diff but never commits: that belongs to the reviewer.
Which CLIs are supported?
There are separate skills for Claude Code, Codex, Cursor, Aider, OpenCode, Cline, Copilot, Kimi, Grok, Qoder and more, with the full list in the README.
Discussion
No questions or reviews yet.
Related
SkillsA CLI that detects your project's stack and installs matching agent skills from a vetted registry in one command
Skills and the porch CLI for agent orchestration: a strong model plans, cheaper ones code, several model families review.
A skill that turns the main Claude Code session into a dispatcher: tasks go to background subagents and the chat never blocks.
A CLI and manager skill: one skills library distributed to 70+ agents, with management done from the agent conversation
A CLI skill manager for AI agents: install from repositories, translate between agent formats and deploy into their configs
A command-line skill manager: installs skills from a marketplace, GitHub, private Git and npm into 45 agents' folders and scans them first
README in short
The README describes a package of delegation skills: one skill per CLI agent plus delegate-setup for building lanes. Each skill carries a single relay.mjs that launches the CLI, writes result.json under the shared delegate-relay.result.v1 contract and reports status, exit code and touched files. It installs through the Skills CLI in one command, for the whole package, for setup only, or for a single skill. The README states plainly what read-only mode each CLI actually enforces. MIT licensed.
README badge
Original SKILL.md
--- name: agy-delegate description: >- Delegate a coding task to the Google Antigravity CLI (`agy`) as a background implementer, then review its diff and land it yourself. Use this whenever the user wants to hand implementation work to Antigravity or agy - phrasings like "have Antigravity do X", "delegate this to agy", "run it through agy", or "use Antigravity to implement/fix/refactor" - or wants to run a queue of coding tasks through agy while staying the reviewer. DO NOT USE for tasks small enough to do inline, or when the user wants the code written directly without delegating. license: MIT compatibility: Requires the `agy` CLI installed and authenticated, Node.js, and git. The orchestrator must be able to run shell commands and read files. Shell examples assume bash/zsh (macOS/Linux, or Git Bash/WSL on Windows). metadata: version: 0.5.0 --- # Antigravity Delegate You are the **orchestrator**. This skill lets you hand a bounded coding task to a separate **implementer** - the Google Antigravity CLI (`agy`) - then 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.