multi-ide-router_skill

This skill intelligently routes tasks between Antigravity IDE and Claude Code CLI to maximize credit efficiency and leverage each environment's strengths.

C#

0

GitHub Stars

1

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill mkalhitti-cloud/universal-or-strategy --skill multi-ide-router

SKILL.md24.4 KB

Overview

This skill intelligently routes development tasks between Claude Code CLI and Antigravity IDE to maximize credit efficiency and leverage each environment’s strengths. It prioritizes Antigravity (Opus Thinking) for any substantive code work and uses Claude Code CLI (Sonnet/Haiku) for planning, context gathering, and routine file operations. The router enforces a confirmation step before generating expensive prompts and includes a clear decision matrix for snippet vs full-file outputs.

How this skill works

The router inspects the task type, complexity, and integration risk, then selects the best IDE and model. For code changes that touch multiple methods or shared state it recommends full-file work in Antigravity Opus Thinking; for isolated edits it recommends snippets via Claude Code CLI Haiku/Opus. It also tracks credit exhaustion and automatically falls back to Claude Code CLI models when Antigravity credits are depleted.

When to use it

Implement new features, refactors, or complex trading logic (use Antigravity Opus Thinking)
Plan work, gather context, and generate prompts for Antigravity (use Claude Code CLI Sonnet)
Run routine file operations, changelogs, or small automated updates (use Haiku sub-agent)
Perform quick fixes or isolated method changes where snippet output is safe (Claude Code CLI)
Fallback to Claude Code CLI when Antigravity credits are exhausted

Best practices

Always answer Sonnet’s clarifying questions before prompt generation
Follow the snippet vs full-file decision matrix: prefer full file for shared state or multi-method changes
Confirm cost vs risk when Sonnet recommends full-file output — user must approve
Use the provided handoff templates to transfer context cleanly between IDEs
Keep live trading changes conservative: prioritize full-file outputs for safety

Example use cases

Designing a new subsystem or major refactor — Antigravity Opus Thinking with full-file output
Live trading emergency diagnosis and mitigation — Antigravity Opus Thinking for deep reasoning
Complex debugging workflow: start in Claude Code CLI, escalate to Antigravity if unresolved
UI layout change that touches one method — generate a snippet and let Haiku integrate it
Implement MOMO trade setup — full-file approach to avoid integration defects

FAQ

The router falls back to Claude Code CLI sub-agents (Opus/Haiku) and queues the task until credits refresh; resume in Antigravity when available.

Who decides snippet vs full-file output?

Sonnet analyzes changes using the decision matrix and recommends an approach; the user must confirm the recommendation before prompt generation.