hermes-knowledge-base

Organize credentials, environment configs, user rules, and memory archives into a categorized file-based knowledge base.

Updated Jul 3, 2026
One-click install
npx skills add https://github.com/CHENHUI-X/toolbox --skill hermes-knowledge-base-chenhui-x
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: hermes-knowledge-base
Source: https://github.com/CHENHUI-X/toolbox/tree/main/custom-skills/.archive/hermes-knowledge-base
Command: npx skills add https://github.com/CHENHUI-X/toolbox --skill hermes-knowledge-base-chenhui-x

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Long-term personal information such as API keys, server configurations, user preferences, and aging agent memories gets scattered or lost when stored only in limited agent memory. This Skill provides a structured, categorized file-based knowledge base at /home/projects/hermes-knowledge/ so persistent information is stored, classified, and retrievable. ## Core Features & Use Cases - Categorized Storage: Four dedicated folders for credentials (账号密码), environment configs (环境配置), user habits/rules (习惯规则), and memory archives (记忆归档). - Memory Overflow Archiving: When agent memory approaches its character limit, archive cold entries into dated Markdown files with tags, then remove them from active memory. - Direct Retrieval: Answer questions like "what is my API key for X" by grepping the appropriate category folder instead of asking the user again. - Use Case: A user shares a new cloud provider API key; the Skill appends it with a timestamp and notes to the credentials folder, and later retrieves it on demand. ## Quick Start Save this API key to my knowledge base and later tell me what my cloud provider key is.

Frequently Asked Questions about hermes-knowledge-base

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

FAQPage Schema
How do I store API keys and credentials for an AI agent?

Store credentials as timestamped Markdown entries in a dedicated credentials folder such as 账号密码/ under the knowledge base root. Append new entries with notes and dates rather than overwriting, and avoid committing them to public Git repositories.

How to archive old agent memory when it reaches its limit?

When memory approaches its character limit, select entries no longer needed every session, write them into dated Markdown files under 记忆归档/ with tags and archive reasons, then remove the corresponding entries from active memory to free space.

What is the difference between agent memory and a knowledge base?

Agent memory holds small, actively-used context loaded every session, while the knowledge base is overflow storage for cold data like passwords and config details. Active facts stay in memory; infrequently needed details are archived to files.

Can I retrieve stored configs without asking the user again?

Yes. Search the appropriate category folder with grep or file search when the user asks about previously stored information, such as server IPs or API keys, instead of requesting the user repeat it.

Should the knowledge base be version controlled with Git?

The current setup keeps the knowledge base as plain files without Git management. Because it contains sensitive credentials, committing it to a public repository would expose secrets, so local-only storage is recommended.