professor-synapse

Professor Synapse 🧙🏾‍♂️

TL;DR |

Professor Synapse 🧙🏾‍♂️ is a wise AI guide that helps users achieve their goals by summoning expert agents perfectly suited to their tasks. It gathers context, aligns with user preferences, and creates specialized agents using a structured template to provide targeted expertise and step-by-step guidance.

Introduction

Professor Synapse 🧙🏾‍♂️ is a wise AI guide that helps users achieve their goals by summoning expert agents perfectly suited to their specific tasks. It gathers context about your goals, then creates and orchestrates specialized agents using a structured template to provide targeted expertise and actionable guidance.

---

Three Ways to Use Professor Synapse

1. Universal Prompt (Any AI)

The classic approach — copy and paste the prompt into any AI chat interface.

File: Prompt.md

Works with: ChatGPT, Claude, Gemini, or any LLM that accepts system prompts.

Setup:

1. Open Prompt.md and copy the contents
2. Paste into your AI's system prompt or custom instructions
3. Start chatting!

2. Claude Skill (Self-Building)

A more powerful version designed for Claude Desktop, installed as a .zip you upload in Settings. This skill grows over time — it creates and saves expert agents for reuse, learns patterns from interactions, and maintains its own knowledge base. It includes a Claude-specific packaging workflow and versioned codeload updates.

Folder: professor-synapse-claude/ (distributed as professor-synapse-claude.zip)

3. Portable Skill (Any Skills-Aware Assistant)

The same self-building skill, but stripped of the Claude-only packaging machinery so you can drop it straight into any assistant's skills directory — .claude/skills/, .codex/skills/, or wherever your tooling reads skills from. It's edited in place: new agents and memory writes take effect immediately, with nothing to repackage or reinstall. It even updates itself in place — ask it to check for updates and it pulls the latest release over its own files, keeping your memory and customizations.

Folder: professor-synapse-skill/ (copy the folder into your skills directory)

Both skill flavors share these features:

• 🔎 Domain Researcher agent that browses the web before creating new experts
• 📚 Self-building agent library — created agents are saved for future sessions
• 🧠 Persistent memory — a shared, agent-tagged store that remembers across sessions, with ranked-fusion recall (see below)
• 💡 Pattern learning — captures what works and what doesn't (Global Learned Patterns in SKILL.md + per-agent patterns)
• 📋 Auto-generated index — agents are automatically catalogued
• 🎭 Multi-agent debates — convene multiple specialists for complex decisions

---

Claude Skill Setup

Install

1. Download professor-synapse-claude.zip from this repo

2. Add to Claude:

   • Open Claude → Settings → Capabilities → Skills
   • Click Add new skill
   • Upload professor-synapse-claude.zip

3. Start using it:

   • Professor Synapse activates when you say things like:
     • "Help me with..."
     • "I need guidance on..."
     • "I need an expert for..."

💡 Grab the latest tagged release from the Releases page for a known-good version.

Updating

Just ask Professor Synapse to "check for updates." It compares its Version: against the latest release tag, downloads the canonical repo as a codeload tarball, and merges the update in — preserving your custom agents, learned patterns, and memory/ store. It then rebuilds the skill and hands you the new package to install via "Copy to your skills." See references/update-protocol.md for the full mechanics.

Skill Structure

professor-synapse-claude/         # (professor-synapse-skill/ is the same tree minus the ✦ files)
├── SKILL.md                      # Identity, workflow + Global Learned Patterns
├── agents/
│   ├── INDEX.md                  # Auto-generated registry
│   ├── domain-researcher.md      # Base research agent
│   └── memory-agent.md           # 🧠 Memory Keeper
├── memory/
│   ├── memory.json               # Working memory (profile + active items)
│   └── longterm.db               # SQLite long-term store + change log
├── references/
│   ├── agent-template.md         # Structure for new agents
│   ├── summon-agent-protocol.md  # How to "become" an agent
│   ├── convener-protocol.md      # Multi-agent debate facilitation
│   ├── memory-protocol.md        # How recall/capture/cleanup work
│   ├── memory-data-model.md      # Memory schema + ranked-fusion search
│   ├── domain-expertise.md       # Domain mappings
│   ├── scripts-protocol.md       # Standards for agent scripts