A standalone build, connected to nothing else in this collection. You drop all your files into a folder; the skill teaches as it works while it extracts the text, generates embeddings with a local model (no API key, and nothing leaves your machine), stores them in an embedded Postgres database with pgvector inside the folder, and draws a Graphify knowledge graph connecting your files by folders, links, and semantic similarity, with an interactive viewer. A simple local MCP server makes the whole thing semantically searchable from any Claude session, and a watcher keeps the embeddings and the graph current whenever you add, change, or remove a file. Along the way it suggests personal ways to use the brain, drawn from your role and what your files reveal.
- Intake: drop everything in the folder; the inventory says what will extract and what will not, honestly.
- Build: extract, embed locally with no API key, and draw the knowledge graph, with the concepts taught as each one becomes real.
- Prove it: the first semantic searches, run together, sources shown.
- Serve and watch: the local MCP server registered with your yes, the watcher keeping everything current, and the guide written into the folder.
You provide
- A folder with your files in it (any formats; subfolders welcome)
- A yes or no on the MCP registration and the watcher, each proposed explicitly
- Your reactions to the personalized usage ideas
You receive
- brain/: embedded Postgres with pgvector, local embeddings, the Graphify graph and viewer, the MCP search server, and the watcher
- second-brain-guide.md: what was indexed, how it works, how to query it, and the usage ideas you kept
Download this skill and its install guide: https://exec-skills.paulcheek.com/skills/cheek-build-a-second-brain
How to use this skill
Once installed, start a new conversation and paste this prompt. The slash command invokes the skill directly; everything after it is your opening brief, so edit it to fit.
Sample output
A completed run of this skill for a sample company, Mastercard, exactly as the deliverable would arrive. Open it in a new tab.
The samples are illustrative demonstrations of the output format, built from public information. Mastercard is not affiliated with and has not endorsed these materials.