SantanderAI/ralph-vault-skill ? reverse-engineered prompt

Reverse engineered prompt

Build me a Python tool and agent skill that helps me create and keep updated a progressive knowledge vault for one or many code repos, like a tiered deepwiki for AI agents.

I want it to let me initialize a vault, add or remove repos and subfolders, check what is missing or stale, validate the docs, and generate a plan that another agent can follow to fill in the vault one repo at a time. It should organize the docs into clear layers like index, repos, components, infrastructure, technologies, relations, and meta, and keep the content driven by fixed prompts so the output stays consistent. Please make it work as a CLI and as a skill that an agent can follow from a simple natural language request.

Use plain Markdown files and a small config file, keep the deterministic parts separate from the AI written parts, and make sure it can handle multiple repos without hardcoding anything. If you need to check current docs or best practices online, go ahead and do that.

Are you gonna build this?

make sure you review the code using coderabbit

Try freeSponsored — opens CodeRabbit in a new tab