cheese

Route arbitrary user inputs to the correct workflow skill via intent classification.

15|1|Updated May 4, 2026
One-click install
npx skills add https://github.com/paulnsorensen/easy-cheese --skill cheese
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: cheese
Source: https://github.com/paulnsorensen/easy-cheese/tree/main/skills/cheese
Command: npx skills add https://github.com/paulnsorensen/easy-cheese --skill cheese

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill solves the problem of figuring out which downstream workflow step to run for an arbitrary user input, and routing that input to the correct skill immediately.

Core Features & Use Cases

  • Unified entry point: Accepts dropped-in content (ideas, spec paths, files, PR/issue refs, stack traces, or simple “/cheese” messages) and turns it into a confident intent-to-target dispatch.
  • Autonomous routing with safety gating: Dispatches immediately by default, but can gate dispatch behind a confirmation prompt with --safe.
  • Escalation when uncertain: Performs silent internal analysis and can ask exactly one clarifying question when needed to avoid misrouting.
  • Pipeline resumption: Supports --continue <slug> to resume from the latest handoff state without requiring the user to re-specify the whole context.

Quick Start

Use the cheese skill to route your request by saying: /cheese add dark mode to the web client.

Frequently Asked Questions about cheese

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

FAQPage Schema
How do I route arbitrary code inputs to the correct workflow?

Route arbitrary code inputs by classifying intent shapes from dropped content like feature requests, file references, or stack traces, immediately dispatching them to the appropriate downstream skill.

Can I gate workflow dispatch behind a confirmation prompt for debugging?

You can gate workflow dispatch for debugging by using the safe flag, which prevents autonomous routing until confirmation, ensuring sensitive debugging material is handled correctly.

What happens when intent classification is uncertain during agent routing?

When intent classification is uncertain during agent routing, it performs silent internal reasoning and asks exactly one clarifying question to prevent misrouting.

How do I resume a workflow pipeline from a previous handoff state?

Resume a workflow pipeline from a previous handoff state by using the continue flag with a slug, restoring the latest context without requiring full re-specification.

Does workflow dispatch work with PR and issue references?

Workflow dispatch works with PR and issue references by parsing them as valid input shapes alongside spec paths, file references, and stack traces for intent classification.

What is the best way to handle intent classification for multiple code review inputs?

The best way to handle intent classification for multiple code review inputs is using a unified entry point that accepts varied content types and turns them into confident intent-to-target dispatches.