impact-check

Assess the importance and reach of a research idea before committing effort.

75|7|Updated May 2, 2026
One-click install
npx skills add https://github.com/zjunlp/Mechanist --skill impact-check-zjunlp
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: impact-check
Source: https://github.com/zjunlp/Mechanist/tree/main/skills/impact-check
Command: npx skills add https://github.com/zjunlp/Mechanist --skill impact-check-zjunlp

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Researchers risk spending months on problems that turn out to be unimportant or low-reach. This Skill evaluates whether a proposed research idea or behavior studies a problem that actually matters, before any experimental commitment. ## Core Features & Use Cases - Impact Dimension Scoring: Rates an idea across five dimensions — important problem, uptake/citation, direction-shifting, real-world reach, and phenomenon value — with HIGH/MEDIUM/LOW judgments. - External Cross-Model Review: Sends the assessment to an external LLM reviewer via the llm-chat MCP for strict, high-rigor feedback and the strongest "so what?" objection. - Structured Impact Report: Produces a report with a 0-10 score, a PROCEED / PROCEED WITH CAUTION / DEPRIORITIZE recommendation, and the audience that would care most. - Use Case: Before launching a mechanistic interpretability experiment, ask whether the target behavior is an open, actively studied problem and whether results would change how the field works. ## Quick Start Ask the assistant to run an impact check on your research idea, for example: "Impact check: studying whether attention heads in small LLMs encode factual recall circuits."

Frequently Asked Questions about impact-check

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

FAQPage Schema
How do I check if my research idea is important before starting?

Run an impact check that scores the idea across five dimensions: problem importance, likely uptake, direction-shifting potential, real-world reach, and phenomenon value. The result is a 0-10 score with a proceed or deprioritize recommendation.

What criteria make a research problem high impact?

A problem is high impact if it solves a needed problem, is likely to be used or cited by follow-up work, could shift a research direction, helps real applications or cross-disciplinary work, or reveals an important phenomenon even with a simple method.

Does the impact check use an external model for review?

Yes, it calls an external LLM reviewer through the llm-chat MCP for strict, high-rigor feedback. The reviewer is asked who would care, how far the result would reach, and the strongest "so what?" objection.

When should I not rely on an impact check alone?

An impact check judges importance, not novelty or feasibility. An idea can be important but already solved, or important but intractable with your resources, so pair it with novelty and feasibility assessments before committing.

What output does the impact check produce?

It produces a structured markdown report with per-dimension HIGH/MEDIUM/LOW ratings, an overall score out of 10, a PROCEED / PROCEED WITH CAUTION / DEPRIORITIZE recommendation, the strongest case for why it matters, and the main objection.