Hermes Agent Masterclass: 10. Security

The finale of my Hermes Agent Masterclass: you've built an agent that can touch files, run shell, spend money, and coordinate a team — this module is how you keep that power from becoming a liability. Sign up for my FREE weekly newsletter, where I spill my unfiltered thoughts on the latest AI news, cool research, and projects I'm building: https://www.onchainaigarage.com/ 🧠 Run agents? Give them a knowledge base. Agent Wikis has free, curated LLM wikis on a ton of AI topics — plus a Pro tier ($9.99/mo) for super-sized wikis with way more depth: https://agentwikis.com/ Module 10 wraps the series with security — the other half of everything the previous nine modules built. The core idea: there's no fully-secure, fully-capable agent, so the seven layers are dials you tune to the trust level you actually have, from a solo laptop to a public-facing gateway. I walk through the trust layer (allow lists so only you can even message it), approvals (manual vs. smart vs. YOLO — and the commands YOLO still refuses), containment (running Hermes inside Docker), and the filters that redact your secrets and block prompt-injection from your own files. I also cover Tirith, the tool that catches sneaky homoglyph and injection attacks, and how to set all of this per-profile. That's defense in depth — and the end of all ten modules. Resources: 🔗 Hermes Agent: https://github.com/NousResearch/herme... Timestamps: 0:00 - Module 10 (finale): security 2:05 - The capability vs security trade-off 2:53 - Where the dials sit (solo / team / public) 5:00 - Trust layer: who can talk to your agent 7:51 - Approvals: manual, smart, and YOLO 11:26 - The YOLO floor (what it still refuses) 13:20 - Containment: running Hermes inside Docker 17:23 - Filters: secret redaction + injection scanning 20:08 - Tirith + per-profile security + series recap Thanks for watching the whole Masterclass — all 10 modules, 5+ hours. Let me know in the comments what you thought of the series and what you'd want covered next. I'll keep exploring Hermes as it evolves, so please like, subscribe, and hit the bell! 🦐✨ #HermesAgent #AISecurity #AIAgents #Docker #AgenticAI #PromptInjection #SelfHosted #AITools #Masterclass