Claude Skill MartBrowse skillsQuick linesLearn by videoTerminal guideWhat is a Skill?
 
← Back to list

Autoskill (turn your workflow history into skills)

Reads your local screenpipe timeline, clusters repeated workflows, and drafts new skills or composition recipes for the patterns not yet covered.

AutomationAdvanced★ 41,485⑂ 3,824AI score 8/10Last updated: Aug 31, 2026

What it does

  • Pulls a time-windowed activity timeline from your locally running screenpipe daemon (localhost:3030) — it has no other data source and refuses to run if the daemon is unreachable.
  • Redacts emails, API keys, bearer tokens and phone numbers, segments sessions on idle gaps, and clusters repeated app-signature workflows entirely on-device.
  • Matches each cluster against the repo's 135 existing skills with local sentence-transformer embeddings, then classifies it as reuse, compose, or novel.
  • Writes proposals to ~/.autoskill/proposed/<timestamp>/ (report.md + draft SKILL.md files) and promotes approved ones into skills/<name>/.

Who it's for

  • Researchers and analysts who repeat the same literature / figure / slide routines daily
  • Power users who want their agent to automate their actual observed workflow
  • People comfortable self-hosting screenpipe plus a local LLM (LM Studio)

Examples

  1. "Analyze my last 4 hours and propose skills that aren't covered yet" → clusters the window and writes report.md
  2. Run with --dry-run to inspect plan.md (the exact timeline to be analyzed) before any LLM call
  3. Happy with a draft? promote --name zotero-pubmed-helper moves it into the skills directory

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/K-Dense-AI/scientific-agent-skills/HEAD/skills/autoskill/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the skills/autoskill folder from the GitHub repo K-Dense-AI/scientific-agent-skills into my ~/.claude/skills/autoskill/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone --depth 1 https://github.com/K-Dense-AI/scientific-agent-skills.git /tmp/sas && mkdir -p ~/.claude/skills && cp -r /tmp/sas/skills/autoskill ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Clone the repo: git clone --depth 1 https://github.com/K-Dense-AI/scientific-agent-skills.git
  2. Install the skill: mkdir -p ~/.claude/skills && cp -r scientific-agent-skills/skills/autoskill ~/.claude/skills/
  3. Install and run screenpipe (mandatory): git clone --depth 1 https://github.com/mediar-ai/screenpipe.git && cd screenpipe && cargo build -p screenpipe-engine --release, then ./target/release/screenpipe record --disable-audio --use-pii-removal. Grant macOS Screen Recording permission when prompted.
  4. Export the token: export SCREENPIPE_TOKEN=$(screenpipe auth token)
  5. Install Python deps: pipenv install httpx pyyaml sentence-transformers (the ~80 MB embedding model downloads on first run).
  6. Set up the default local LLM: install LM Studio and run lms load gemma-4-31b-it --gpu max -y; for cloud, flip backend in config.yaml and set ANTHROPIC_API_KEY or FOUNDRY_API_KEY.
  7. Verify with python scripts/autoskill.py doctor --config config.yaml --skills-dir ../, then ask Claude Code: "analyze my recent work and propose skills."
View source on GitHub ↗License: MIT