Back to the catalog

ice-framework

The ICE Plugin — an open-source Claude Code plugin packaging the ICE planning framework (Intent, Constraints, Expectations) as a distributab

Open source Repository Open in the app JSON README (API)

About

The ICE Plugin — an open-source Claude Code plugin packaging the ICE planning framework (Intent, Constraints, Expectations) as a distributable skill for the Claude Code marketplace. ICE structures planning conversations between human and AI before implementation begins, surfacing requirements through structured dialogue, enforcing critical review, and producing a documented blueprint. It reduces agreement bias and vague-input failures. The plugin packages the following skills: /ice — Full ICE planning session with blueprint output /q — Clarifying questions before starting work /challenge — Challenge assumptions from hostile positions /verify — Sceptical review of AI output before accepting

Details

Kind
Plugins
Topic
No topic detected
Publisher
topsy31
Origin
marketplace
Category
ferramentas
Last push
2026-04-16T13:22:29Z
Repository state
ativo
Added
2026-08-30 01:48:58
Updated
2026-08-30 01:48:58
Origin id
topsy31/ice/ice-framework

README

# ICE Framework — Claude Code Plugin

A Claude Code plugin packaging the ICE planning framework: a structured dialogue system that surfaces requirements, enforces critical review, and produces documented blueprints before implementation begins.

## Skills

| Skill | Purpose |
|-------|---------|
| `/ice-framework:ice` | Full ICE planning session — Intent, Constraints, Expectations — with blueprint output |
| `/ice-framework:q` | Clarifying questions before starting work |
| `/ice-framework:challenge` | Challenge assumptions from hostile positions |
| `/ice-framework:verify` | Sceptical review of AI output before accepting |

## Why ICE?

Most poor AI results stem from vague inputs, not AI limitations. The ICE framework front-loads the thinking that should happen before work begins:

- **Intent** — What are we actually trying to achieve, and why?
- **Constraints** — What limits exist? What can we not do?
- **Expectations** — How will we verify success?

The `/challenge` and `/verify` skills form a critical-thinking chain around `/ice`, prompting the plan to be stress-tested before building and the output to be reviewed carefully after.

## Install

Add the marketplace from GitHub, then install the plugin:

```
/plugin marketplace add Topsy31/ICE
/plugin install ice-framework@ice-framework
/reload-plugins
```

Skills are then available as `/ice-framework:ice`, `/ice-framework:q`, etc.

If you prefer shorter names, you can also install directly from the official Anthropic marketplace once the submission is accepted:

```
/plugin install ice-framework@claude-plugins-official
```

To submit or check submission status: [platform.claude.com/plugins/submit](https://platform.claude.com/plugins/submit)

## Usage

Run before starting any complex or ambiguous task:

```
/ice-framework:q        ← Start here to clarify requirements
/ice-framework:challenge ← Challenge your approach
/ice-framework:ice       ← Full planning session with blueprint
/ice-framework:verify    ← Review AI output before accepting
```

### Choosing the right skill

| Situation | Start with |
|-----------|-----------|
| You have a rough idea and aren't sure what you want | `/ice-framework:ice` — full planning session with blueprint |
| You have a specific request but want to ensure nothing's missed | `/ice-framework:q` — targeted clarifying questions before work begins |
| You have a plan and want it challenged before building | `/ice-framework:challenge` — hostile-position critique of the approach |
| Work is done and you want to review AI output before acting on it | `/ice-framework:verify` — structured checklist for human review |

### Recommended workflow for complex tasks

```
/ice-framework:q  →  /ice-framework:challenge  →  /ice-framework:ice (blueprint)  →  build  →  /ice-framework:verify
```

For simpler tasks, use only what adds value. `/ice-framework:q` alone is enough for a focused request with a clear scope.

## When to use ICE

ICE adds value for:
- Complex, ambiguous tasks
- Multi-step implementations
- Tasks where requirements might evolve
- Anything where getting it wrong would be costly

ICE is overkill for:
- Simple, well-defined tasks
- Quick fixes with obvious solutions

## Licence

MIT

More