canaryclaude-cheap-subagent

Route mechanical coding edits to cheaper subagents via canaryclaude.

17|1|Updated Feb 2, 2024
One-click install
npx skills add https://github.com/FormalSnake/dotfiles --skill canaryclaude-cheap-subagent
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: canaryclaude-cheap-subagent
Source: https://github.com/FormalSnake/dotfiles/tree/main/users/kyandesutter/claude/skills/canaryclaude-cheap-subagent
Command: npx skills add https://github.com/FormalSnake/dotfiles --skill canaryclaude-cheap-subagent

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill solves the problem of spending time on repetitive, mechanical coding edits by letting you delegate the typing-heavy parts while you handle the reasoning.

Core Features & Use Cases

  • Cheap parallel subagent routing: Offloads many near-identical refactors or formatting passes to canaryclaude so you can run lots of edits concurrently.
  • Model-specific prompt discipline: Helps you target fast/cheap models for mechanical work and tighten prompts when using local Qwen (including forcing explicit write tool invocation).
  • Autonomous editing with guardrails: Uses a read-then-write workflow, requires precise “what to change” and “what not to do,” and emphasizes verification via git diff and tests.

Quick Start

Use canaryclaude to apply an explicit, numbered refactor spec to a specific file path, then verify the batch with git diff and run the relevant tests.

Frequently Asked Questions about canaryclaude-cheap-subagent

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate repetitive code refactoring across multiple files in Claude Code?

Automating repetitive code refactoring is handled by offloading mechanical edits to a cheaper subagent. This allows you to execute near-identical transformations like renames and formatting concurrently across many files.

What is the best way to parallelize mechanical code edits and boilerplate generation?

Parallelizing mechanical code edits requires dispatching execution-heavy, reasoning-light tasks to a cheaper model. This approach uses a read-then-write prompt workflow to apply fixed patterns per file autonomously.

How do I run autonomous file editing with guardrails to prevent unwanted changes?

Autonomous file editing uses a permission-mode bypass combined with strict prompt discipline. You must specify exactly what to change and what not to do, mandating verification via git diff and tests afterward.

Can I use local Qwen models for automated code refactoring and formatting?

Yes, you can use local Qwen models for automated code refactoring by targeting fast models for mechanical work. This requires tightening prompts and forcing explicit write tool invocation during the editing process.

Does dispatching cheap subagents for code edits require explicit model selection?

Yes, dispatching cheap subagents requires explicit model selection alongside a fish-based invocation through a CanaryLLM proxy. This setup ensures execution-heavy transformations are routed to the correct model tier.