autoskill

Cluster repeated screenpipe workflows and propose reusable skill drafts.

22|4|Updated May 25, 2026

Search Tags:#automation #workflow #prototyping #skill-generation #local-llm #screenpipe

Ready to run

npx skills add https://github.com/crazymsn/academic-skills --skill autoskill-crazymsn

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: autoskill
Source: https://github.com/crazymsn/academic-skills/tree/main/academic-skills/autoskill
Command: npx skills add https://github.com/crazymsn/academic-skills --skill autoskill-crazymsn

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires httpx, pyyaml, sentence-transformers, and includes scripts (resource) and references (resource) components.

What problem does it solve?

Analyzes local workstation activity to automatically identify repetitive research workflows and convert them into reusable skills that extend your automation library.

Core Features & Use Cases

Clusters repeating user workflows from screenpipe data and matches them against existing skills to propose new ones or composition recipes.
Generates staged SKILL.md drafts that you can review and promote into academic-skills/, enabling rapid expansion of capabilities.
Keeps all processing local, redacts sensitive information before any draft is produced, and supports offline use with a local LM backend.

Quick Start

Run autoskill with a start and end time window to generate draft proposals from your recent workflows.

Frequently Asked Questions about autoskill

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

FAQPage Schema

How do I convert repeated local workflows into reusable automation skills?▼

To convert repeated local workflows into reusable skills, you can analyze screenpipe activity data to identify patterns, match them against an existing skill library, and generate staged SKILL.md draft proposals for review.

Do I need a screenpipe daemon running to identify workflow patterns from local activity?▼

Yes, identifying workflow patterns requires a screenpipe daemon running on localhost:3030 to provide local activity data. You also need a local LM backend and a writable academic-skills repository to store the generated proposals.

How does local workflow analysis handle sensitive data during skill generation?▼

Local workflow analysis handles sensitive data by performing all processing on-device and applying redaction to sensitive information before any skill proposal drafts are generated, ensuring privacy and supporting offline use.

Can I generate skill composition recipes by matching screenpipe activity to an existing skill library?▼

Yes, you can generate skill composition recipes by clustering repeating user workflows from screenpipe activity and matching those patterns against your existing academic-skills library to propose new skills or composition recipes.

How do I create draft skill proposals from a specific time window of workstation activity?▼

You create draft skill proposals by running the analysis tool with a designated start and end time window, which extracts recent workflow data from screenpipe, clusters the activity, and outputs staged SKILL.md drafts for promotion.