claude-orator — Claude Code plugin: prompts and installation
claude-orator is a Claude Code plugin that scores and restructures prompts before they go to subagents. Its hook suggests a rewrite for underspecified Task prompts.
MIT ★ 78 updated 2026-03-05
/plugin marketplace add Vvkmnn/claude-emporium
/plugin install claude-orator@claude-emporiumWhat claude-orator does
claude-orator scores prompts across seven dimensions and returns a restructured version. Its MCP server selects applicable techniques from Anthropic’s recommendations. It makes no LLM calls or network requests; processing stays in memory.
The PreToolUse(Task) hook checks prompts longer than 50 characters. If one has no XML tags, Markdown headings, or action verb, the hook suggests running orator_optimize. It never blocks the call. The plugin requires the claude-orator-mcp server, added with claude mcp add orator -- npx claude-orator-mcp.
What it helps with
- score a prompt for clarity and structure
- rework a prompt before handing it to a subagent
- flag an underspecified prompt during a
Taskcall
Example prompts for the agent
- Rewrite this prompt for a subagent: “Check the payment handler”
- Restructure my request to include clear constraints and context
- Score this prompt with
/reprompt-orator
How to install claude-orator
Plugin marketplace
/plugin marketplace add Vvkmnn/claude-emporium/plugin install claude-orator@claude-emporiumStep by step, with commands for each platform: Claude Code plugins: marketplaces, installing, and building your own
Questions about claude-orator
How do I install claude-orator in Claude Code?
Run these commands in order: /plugin marketplace add Vvkmnn/claude-emporium, then /plugin install claude-orator@claude-emporium.
What does claude-orator do?
It scores a prompt across seven dimensions and returns a restructured version. Its Task hook can suggest an optimization for an underspecified subagent prompt.
What does claude-orator need to run?
It requires the claude-orator-mcp server. Add it with claude mcp add orator -- npx claude-orator-mcp.
Source text
claude-orator · MIT
# Claude Orator
Prompt rhetoric coach for Claude Code. Deterministically scores and restructures prompts using Anthropic best practices. No LLM calls, no network, in-memory only.
## Installation
```bash
/plugin marketplace add Vvkmnn/claude-emporium
/plugin install claude-orator@claude-emporium
```
## Requirements
MCP server: [`claude-orator-mcp`](https://www.npmjs.com/package/claude-orator-mcp)
```bash
claude mcp add orator -- npx claude-orator-mcp
```
## Hooks
| Event | Trigger | What It Does |
|-------|---------|-------------|
| `PreToolUse` | `Task` | Suggests optimization for under-specified subagent prompts |
## Commands
| Command | Description |
|---------|-------------|
| `/reprompt-orator <prompt>` | Optimize a prompt using Anthropic best practices |
## MCP Tools
| Tool | Purpose |
|------|---------|
| `orator_optimize` | Score across 7 dimensions, apply techniques, return restructured prompt |
## Scoring Dimensions
| Dimension | What It Measures |
|-----------|-----------------|
| Clarity | Unambiguous language and intent |
| Specificity | Concrete details vs vague requests |
| Structure | Logical organization (XML, headers, sections) |
| Context | Background information provided |
| Examples | Few-shot examples or expected outputs |
| Constraints | Boundaries and requirements stated |
| Tone | Appropriate voice and register |
## Techniques
System prompts · XML tags · Chain of thought · Few-shot · Prefill · Long context · Extended thinking · Tool use
## Settings
Configure in `~/.claude/settings.json` under `claude-emporium`:
```json
{
"claude-emporium": {
"claude-orator": {
"optimize_subagent_prompts": false
}
}
}
```
| Key | Default | What It Controls |
|-----|---------|-----------------|
| `optimize_subagent_prompts` | `true` | Auto-suggest optimization for under-specified subagent prompts |
## How It Works
The hook fires on `PreToolUse(Task)` and runs a quick heuristic: if the prompt is >50 chars and lacks XML tags, markdown headers, or an action verb, it emits a suggestion to run `orator_optimize`. Never blocks — suggestion only.
The MCP server scores deterministically across 7 dimensions (1-10 each), selects applicable techniques from 8 Anthropic best practices, and returns a restructured prompt.
## License
MIT
© vvkmnn · MIT · Vvkmnn/claude-emporium