AI Agent Workflow Collection

Use when Jared asks to build a video content strategy using an AI video generation model (Seedance, Kling, Sora, Veo, etc.). Covers deep model research, brand-aligned prompt library construction, 30-day production calendars, Excel operational plans, and the integration of AI-generated video with real-content companion pieces.

Use when Jared asks for an Apify or Appify Google Maps dashboard, local lead dashboard, niche market scrape, business opportunity dashboard, or signal engine built from Google Maps plus website and enrichment data.

Use when managing Apple apps and device workflows from Hermes on macOS: Notes, Reminders, iMessage/SMS, Find My, and safe background desktop automation.

Use when creating ASCII art, terminal-style diagrams, image-to-ASCII conversions, or full ASCII video/GIF/MP4 treatments.

Use when writing songs, generating music or sound with AI, preparing Suno/HeartMuLa prompts, or analyzing audio features and spectrograms.

Use when delegating coding, PR review, or implementation work to external coding-agent CLIs such as Claude Code, Codex, or OpenCode from Hermes.

Use when Jared asks to "build an agent", "create a new agent", "make me a [role] agent", "design a soul for...", "I need an agent that...", or describes new agent functionality. Covers both Claude Code task agents (frontmatter + system prompt) AND PerformOS specialist souls (identity, brand, voice, domain knowledge). Brock executes the full design — qualified question, format decision, comprehensive soul or agent file, and handoff to deployment.

Use when the user asks to build, deploy, create a dashboard, create a deck, push to GitHub, ship a page/tool, or deploy to Vercel. Orchestrate the Express build protocol using Claude Code, GitHub, and Vercel.

Connects Hermes to Claude OS — the operator dashboard at localhost:9119. Read sessions, memory, integrations, kanban, dream history.

Use when Jared and Brock are upgrading Crew skills to gold-standard depth. Covers shallow upgrades, migration from .claude/skills, merged-source skills, and commodity pack upgrades. Includes prompt templates, Brock review checklist, and architectural conventions.

Cost-controlled AWS EC2 GPU trials for Ollama local-model inference with Hermes agents.

When Jared asks to explain a technical concept, system, or agent workflow in plain English for someone who is not technical (family, clients, executives). Strip jargon, use real-world analogies, and structure explanations so a novice can follow and relay the information to others.

Build premium, self-contained, branded HTML slide decks with smooth animations, navigation controls, responsive layouts, and zero overflow. Use whenever Jared asks for a deck, slides, presentation, HTML presentation, pitch deck, product demo, or branded multi-slide visual output.

Use when Jared asks Brock to build a training programme via LaraLearningDesign. Standard 2-pass delegation pattern to prevent Lara timeout on large builds.

Build LearnOS-ready training module PDFs and Manager Pack companions from a markdown-source pipeline. Handles variable module counts, coloured callout boxes (SHOW, REFLECT, QUIZ, ROLEPLAY, KEY TAKEAWAYS), and exports two PDFs per programme.

Use when discovering, downloading, quantizing, serving, benchmarking, or operating LLMs locally or on GPU infrastructure with Hugging Face, llama.cpp, vLLM, W&B, or lm-eval.

Use when installing, running, importing, exporting, or debugging Jared's local n8n instance and workflows on macOS.

Import, export, and debug n8n workflows on Jared's local Mac instance (port 5678). Covers CLI import format requirements, REST API auth gap, and credential reconnection.

Install, configure, and uninstall NVIDIA NemoClaw with Hermes Agent sandboxes on macOS. Covers provider selection, non-interactive onboarding, Docker prerequisites, common pitfalls, and the Brev cloud escape hatch.

Save Hermes agent conversations into Jared's Obsidian vault under Chat History. Use whenever Jared asks to save, archive, remember, export, or preserve a chat history, and at the end of meaningful agent conversations when practical.

Use when Jared asks about selling or positioning the PerformOS agent package to clients. Covers cloud (Orgo) and local editions, two-tier Standard/Advanced pricing, "up to 10 fully customisable agents" positioning, security framing, compliance checklist, proposal structure, and internal financial model.

How Brock produces PerformOS website builds — research, legal, strategy, markdown deliverables. NEVER deploys. NEVER pushes Vercel or GitHub. Hands markdown files to Jared directly.

Use when turning technical, academic, corporate, or AI-sounding material into clear human language for non-technical readers, clients, frontline staff, or training audiences.

Build airtight Recognition of Prior Learning (RPL) applications that map professional experience to university learning outcomes. Use when the user asks for RPL, prior learning recognition, credit transfer based on work experience, or mapping their job to academic unit learning outcomes.