prompt-improver-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| ANTHROPIC_API_KEY | No | Turns on the Claude API backend. Not set by default. Also supports ANTHROPIC_AUTH_TOKEN or an `ant auth login` profile. | |
| PROMPT_IMPROVER_MODEL | No | The Claude model used for rewriting. | claude-opus-5-5 |
| PROMPT_IMPROVER_EFFORT | No | Reasoning effort for the API call: `low`, `medium`, `high`, `xhigh`, or `max`. | medium |
| PROMPT_IMPROVER_BACKEND | No | Force `anthropic`, `sampling`, or `host`. Set `host` to make sure no API calls are made. | auto |
| PROMPT_IMPROVER_FALLBACKS | No | `0` turns off the server-side retry on a fallback model after a refusal. | 1 |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| improve_promptA | Rewrite a prompt so a model answers it better, using techniques from the prompting literature: LLaVA-CoT staged reasoning, System 2 Attention context distillation, conversational query rewriting, Rephrase-and-Respond, Chain-of-Verification, Step-Back, Plan-and-Solve, Least-to-Most, and more. The server cannot see the conversation, so pass the current prompt verbatim plus the earlier turns and context it depends on. Returns the improved prompt, the techniques used (with paper citations), what changed, assumptions, and clarifying questions. If backend_used is 'host', improved_prompt is a template draft: finish it by following host_instructions. |
| analyze_promptA | Profile a prompt's task type and list the techniques improve_prompt would apply, with citations, without rewriting anything. |
| list_techniquesA | List every technique the improver knows, with the papers behind it. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| improve | Improve a prompt using this conversation as context, then offer to run it. Pass the prompt as the argument, or leave it empty to improve your most recent message. |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 3 tools
improve_prompt (rewrites) and analyze_prompt (profiles without rewriting) are clearly distinct, and list_techniques serves a global reference role. However, analyze_prompt's output (techniques improve_prompt would apply) partially overlaps with list_techniques' output, so the two informational tools could be mildly confused.
All three tools follow a strict verb_noun snake_case pattern: improve_prompt, analyze_prompt, list_techniques. The convention is predictable and readable throughout.
Three focused tools perfectly match the narrow purpose of a prompt-improvement server; each covers a distinct facet (rewrite, analyze, reference) with no filler or redundancy.
The surface covers the core lifecycle of improving, analyzing, and discovering techniques for a prompt. Minor gaps exist: no way to apply/configure specific techniques selectively or manage a history of improved prompts, but these are workable omissions.