Hermes Agent Tutorials & Use Cases

Currently Playing: Trying Out Matt Pocock's /grill-me Skill in Hermes Agent

Can 71 questions save you weeks of building the wrong thing? I tested Matt Pocock's popular /grill-me agent skill on a real game idea to see whether relentless alignment actually improves the plan. 🧠 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/ 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/ Matt Pocock's /grill-me skill is designed to stop an agent from racing into implementation before you agree on what you're actually building. I install it in Hermes Agent, running GPT-5.6 Sol, and use it to stress-test Frontier Labs: my early idea for a SimCity-style strategy game where you run an AI lab, balance compute, training, marketing, safety, reputation, and profit, and learn how AI models work along the way. Across 71 questions and 42 minutes, the skill pushes me to define the audience, win condition, campaign structure, resource system, model-development pipeline, visual direction, first playable vertical slice, tech stack, and business model. It catches contradictions between earlier answers, challenges me when I try to over-scope the MVP, and refuses to implement anything until we confirm a shared understanding. Finally, I turn the decisions into a detailed implementation plan and give my verdict on whether the process was worth the time. Resources: 🔗 Matt Pocock's Skills repo: https://github.com/mattpocock/skills 🔗 Matt Pocock / AI Hero: https://www.aihero.dev/ 🔗 Hermes Agent: https://github.com/NousResearch/hermes-agent 🔗 Agent Wikis: https://agentwikis.com/ Timestamps: 0:00 - Matt Pocock's popular /grill-me agent skill 2:06 - /grill-me vs /grill-with-docs 3:10 - Installing the skill in Hermes Agent 3:25 - Frontier Labs: my AI lab simulation game 4:47 - The grilling begins 7:46 - Ten questions in: refining the gameplay 8:58 - Platform, visuals, MVP, and tech stack 11:58 - Catching contradictions across earlier answers 12:41 - Question 70 and the product model 13:41 - Ending the grill after 71 questions 15:01 - Why no plan or code was created automatically 16:08 - Saving the decisions as an implementation plan 17:08 - Final verdict: is /grill-me worth using? #AgentSkills #GrillMe #MattPocock #HermesAgent #AIAgents #AICoding #AIEngineering #ProductPlanning #GameDev