This repository manages the onboarding documents and protocols for GPT Features Sandbox threads.
These threads are designed for:
- Exploration: Testing GPT capabilities in controlled, documented environments
- Skill Building: Developing user practices around memory handling, feedback loops, and technical collaboration
- Creative Experimentation: Includes technical and narrative testing in areas like system design, writing support, and structured applications
The onboarding document (Gpt Features Onboarding.md) provides a full outline of expectations, flagging protocols, memory management procedures, and continuity strategies.
-
Upload or paste onboarding document at thread start:
- Always paste early in a new thread to ground behavior.
- Re-paste if the session context is lost (e.g., daily resets).
-
Conduct experimental or operational sandboxing:
- Test new model behaviors (marking, memory, thread management).
- Document or discuss changes as needed directly within the thread.
-
Periodically update this repository:
- Download evolved onboarding documents from active threads.
- Apply versioning manually (e.g., v1.0 → v1.1) if significant changes are made.
- Maintain a clear changelog if needed.
Use a .gitignore to prevent clutter from exported chat logs, local scratchpads, and temporary files. A minimal example is provided separately.
- Confidentiality First: Assume all sandbox explorations are personal unless confirmed otherwise.
- Transparency Expected: Uphold marking guidelines within repository-managed threads.
If the GPT Features project scales, future expansions could include:
- Sandbox lineage tracking
- Multi-sandbox comparison tools
- External tooling integrations (e.g., Markdown-to-Git automations)
🌱 This project is a living organism—growing and adapting alongside your collaboration.