AUTOMATION

Bootstrap a Custom Knowledge-Base System with Seed Prompt

Use a single seed prompt to create a markdown-to-HTML knowledge base system with agent interrogation

Updated: 5/22/2026
Difficulty
easy
Time
15m
Use Case
Building a personal, customizable knowledge management system that agents can query and extend without polluting main project context
Popularity
0 views

About this automation

A seed prompt that instructs Claude to generate a Python script converting markdown folders into browsable HTML, creating a flexible knowledge base accessible to agents. The system separates knowledge management from main coding workflows to prevent context memory pollution.

How to implement

1

Create a fresh empty directory (do not use existing project folders)

2

Paste the seed prompt into Claude Code or Cowork project

3

Let the agent onboard you through Phase 1-2 setup

4

Review Phase 3 instructions (10 lines) that build the Python HTML conversion script

5

Approve the script execution when prompted

6

Query and extend the knowledge base via agent interaction

7

Clone additional coding repos into the directory if needed for agent access