Generate Your AI Second Brain Build Plan1 skill
A Claude Code skill that generates personalized PRDs for AI second brains from a requirements template.
AI Agents Blueprint
A skill for Claude Code that generates a personalized PRD for building an AI second brain with memory, integrations, proactive monitoring, and security hardening.
Run this command to deploy the blueprint to your environment.
Builds a complete, phase-by-phase PRD for constructing an AI second brain tailored to your tools and workflow. Instead of starting from a blank template or copying a generic architecture, this skill reads your requirements and generates a personalized build plan that covers memory setup, integrations, proactive systems, and security accounting for your specific platform stack and preferences.
1. Fill out a simple requirements template with your tools, top tasks, platforms, and preferences. 2. Run `/create-second-brain-prd ./my-second-brain-requirements.md` in Claude Code. 3. The skill reads your requirements, researches all APIs and SDKs in your stack via web search, and generates a 9-phase PRD with personalized implementation notes. Output: a build plan at `.agent/plans/second-brain-prd.md` with phase-by-phase instructions, file paths, complexity estimates, and technical specifics for each platform you selected.
A Claude Code skill that generates personalized PRDs for AI second brains from a requirements template.
Every phase contains implementation-ready details for your exact platform stack, eliminating guesswork about API patterns, auth flows, and rate limits.
Requires Claude Code (Claude's built-in code execution environment) and Python 3.9+ for the generated scripts. The generated second brain uses FastEmbed (ONNX model, ~80MB locally) for embeddings, SQLite or Postgres for vector and keyword search, and the Claude Agent SDK for background processes. All generated code assumes Unix-like file paths; Windows users must run under WSL2 or use Windows-native paths carefully.
This generates a PRD, not the actual second brain code. You still need to implement each phase. The generated PRD assumes you can write Python scripts and set up Claude Code hooks; if you need hand-held implementation, pair this with the Dynamous community workshop. Integration setup (OAuth flows, bot tokens) varies by platform and may require manual permission configuration.
Personalized PRD Generation
Reads your requirements (platforms, tasks, proactivity level, security boundaries) and generates a phased build plan with your vault folder name, chosen platforms, and specific technical details.
API Research Integration
Researches every tool and API in your stack (Gmail, Slack, Linear, Asana, etc.) to ensure PRD contains accurate setup, SDK patterns, rate limits, and platform-specific gotchas.
9-Phase Architecture
Covers memory layer, hooks, search, integrations, skills, proactive systems, chat, security hardening, and deployment with dependencies and complexity estimates.
Starter Templates
Ships SOUL.md, USER.md, MEMORY.md, BOOTSTRAP.md (first-run onboarding), and memory folder structure so you don't start from a blank page.
Security-First Design
PRD includes pre-flight guardrails, credential protection hooks, sanitization layers, and dangerous command pattern blockers to secure the system from the start.