Hermes + Discord LLM Wiki Operations System

# Hermes + Discord LLM Wiki Operations System

AI Summary

Purpose:

• Capture durable knowledge about the user's personal AI operations workflow that uses Hermes Agent, Discord, and llm-wiki as a git-backed knowledge base.

Key points:

• Discord is used as the conversational control surface for Hermes Agent.
• Hermes can perform file, git, remote host, browser, scheduled job, and documentation tasks from the chat interface.
• The durable storage target is ~/workspace/repos/llm-wiki, with AI-facing knowledge under ai/ and Korean human-facing outputs under human/.
• The system intentionally does not preserve every chat transcript. It keeps decisions, verification outputs, backup paths, workflow rules, repo notes, and reusable knowledge.
• The workflow already includes conservative git-backed maintenance patterns and hourly sync intent for llm-wiki.

Relevant when:

• Writing portfolio or resume material about personal AI workflow, agent operations, knowledge management, or documentation automation.
• Deciding where to store future conversation/work documentation.

Do not read full document unless:

• You need exact portfolio wording, system boundaries, or what should/should not be claimed.

Linked documents:

• AGENTS.md
• ai/wiki/concepts/agent-work-documentation-policy.md
• ai/wiki/concepts/llm-wiki.md
• ai/repo-notes/hyunwook-tech.md
• ai/worklog/2026/2026-W24.md

Open Questions

• Exact public name for the system: Needs confirmation.
• Whether to expose any implementation screenshots or Discord screenshots publicly: Needs confirmation.
• Usage metrics such as number of saved notes, automated runs, or sessions handled: Needs confirmation.

Details

Project role

This is a personal AI operations system rather than a single-purpose chatbot. The main value is the operating workflow: the user talks to Hermes through Discord, Hermes performs or verifies work with tools, and the results that should survive the conversation are written back into llm-wiki in structured Markdown.

For portfolio copy, frame it as a personal knowledge/work documentation system. Avoid claiming that a new foundation model, generic AI platform, or full autonomous operating system was built.

Operating workflow

1. User starts work from Discord.
2. Hermes performs the requested work using available tools such as filesystem, git, SSH, browser, scheduled jobs, and messaging integrations.
3. Durable outcomes are separated from transient conversation:

- decisions; - commands and verification outputs; - backup and rollback paths; - repo-specific notes; - worklog entries; - reusable policies or concepts.

1. Durable content is committed to llm-wiki, while actual code changes stay in their target development repositories.

Boundaries

What is verified:

• llm-wiki is the central workspace and uses AGENTS.md as the canonical agent instruction file.
• Human-facing outputs under human/** should be Korean and grounded in ai/**.
• Development work should be logged under ai/worklog/ and durable repo knowledge under ai/repo-notes/.
• The user wants meaningful work documentation saved in ~/workspace/repos/llm-wiki.

What should not be claimed without confirmation:

• Enterprise usage.
• Public user adoption.
• Fully automatic judgment without human review.
• Quantified productivity improvements.

Portfolio positioning

Suggested Korean title:

• "Hermes + Discord 기반 개인 LLM Wiki 운영 시스템"

Suggested angle:

• Not "I made a Discord bot", but "I designed a personal AI work interface that turns useful work context into a maintained git-backed knowledge base."

Suggested stack labels:

• Hermes Agent
• Discord
• Markdown
• Git
• Cron
• SSH