How does it research my project?

The skill uses tools like grep, glob, and git log to find relevant code patterns, recent changes, and potential error sources related to your request.

Does it support external research?

Yes, it can use WebSearch and WebFetch to gather information on best practices and online documentation if the local context is insufficient.

What does the Prompt Improver skill do?

It analyzes ambiguous prompts and performs automated research within your codebase to ask clarifying questions before attempting to execute a task.

Can I manually invoke the Prompt Improver?

Yes, you can manually invoke it to enrich a prompt with research-based questions or when building complex evaluation systems.

When is this skill automatically triggered?

It triggers when a submitted prompt is identified as vague, missing specific targets, or lacking sufficient context to be completed accurately.

Prompt Improver

Name: Prompt Improver
Author: Activer007

byActiver007

0•

Productivity & Workflow

Transforms vague user prompts into actionable development tasks through systematic codebase research and targeted clarification questions.

The Prompt Improver skill acts as an intelligent bridge between ambiguous user requests and high-precision execution. When a prompt is identified as lacking specific context or targets, this skill automatically initiates a multi-phase workflow: it researches the local codebase using grep and git logs, reviews conversation history, and identifies relevant documentation. Based on these findings, it generates grounded multiple-choice questions to resolve uncertainty, ensuring that Claude never acts on assumptions and always delivers accurate results aligned with the project's actual structure.

Key Features

01Deep codebase exploration including file patterns and git history

020 GitHub stars

03Contextual awareness of conversation history to avoid redundancy

04Automatic enrichment of ambiguous or under-specified prompts

05Research-grounded multiple-choice question generation

06Seamless transition from clarification to task execution

Use Cases

01Refining architectural changes by presenting research-backed implementation options

02Resolving high-level requests like 'fix the bug' by identifying specific error locations

03Onboarding to unfamiliar repositories by bridging the gap between intent and file structure

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add activer007/ordinary-claude-skills prompt-improver

For use in Claude.ai and ChatGPT

Download Skill