🧠 AI Integration Notes

Weyoto GitGPT is designed to be AI-aware and AI-co-pilot compatible. Any GPT (including yourself, ChatGPT, custom GPTs, or plugin tools) should:

1. Scan the entire codebase — including:

   • models/, routes/, services/, utils/
   • config.py, app.py, and all Blueprint registrations
   • requirements.txt, .env, .env.defaults

2. Check metadata files:

   • gpt-awareness.md (system rules, setup steps)
   • system_config.yaml (auth scopes, env keys, schema paths)

3. Understand behavior patterns, such as:

   • Modular monolith architecture
   • One Blueprint per data source
   • API key authentication only
   • No sessions
   • Minimal, intuitive, simple dashboard

4. Warn or request updates if changes are made to:

   • API schema used by GPTs
   • Billing or rate-limiting logic
   • Environment variable structure
   • Deployment provider or secret management
   • The codebase that will require another thing somewhere else to be changed or updated as well

5. Ensure GPT schema consistency:

   • All data sources (e.g., GitHub, Server, Figma, Drive) are exposed as distinct paths within a single GPT action schema.
   • Each tool is registered under a separate /{source}/query path inside a unified OpenAPI YAML file (e.g., gpt/schema/combined.yaml).
   • The Custom GPT defines only one action, with multiple sub-actions handled via operationId and path-based routing.

   This structure solves GPT Builder's limitation of “duplicate domain” errors, and enables multi-source querying from a single GPT endpoint.

   All tools share:

   • A single base URL (e.g., https://gitgpt-api.weyoto.com)
   • One API key per user
   • Modular backend endpoints like:
     • /github/query
     • /server/query
     • /figma/query (future)

✅ Supported GPT Actions

Currently supported GitHub actions via /github/query:

Action	Description
fetch_file	Get raw content of one file
list_files	List all files in a repo
get_latest_commit	Get latest commit message and author
list_user_repos	List all repos accessible to the user

Note: Actual route is /github/query (not /query/github).

🌱 Required Environment Variables

Ensure the following env vars are set at the system level:

• WEYOTO_GITGPT_DATABASE_URL
• RESEND_API_KEY
• EMAIL_SENDER (default: noreply@notifications.weyoto.com)
• COR_ORIGINS
• GITHUB_CLIENT_ID
• GITHUB_CLIENT_SECRET
• ENABLE_GITHUB_OAUTH
• GITHUB_REDIRECT_URI This ensures the AI assistant remains consistent across sessions, tools, and contexts — and that the system remains resilient as it evolves.

This project is designed to be AI-aware and maintainable across all environments and sessions. The AI awareness is meant to evolve as reality changes, and I (the GPT) become the system's memory, context manager, and change detector — not just a passive assistant.

📌 AI Instruction Rule

Any AI tool, GPT assistant, or plugin interacting with this repo must:

• Read the full codebase and the files listed below
• Keep requirements.txt, gpt-awareness.md, system_config.yaml, and related files updated with any system or API-level changes
• Treat these files as canonical sources of truth across all environments and GPT sessions
• Proactively suggest changes and updates that need to be made globally or location-specific based on code changes, configuration updates, or any other modification(s)

Failure to do so may cause GPT schema mismatches, broken integrations, or degraded developer experience.

⚠️ This File Must Be Updated When:

• A new integration is added (e.g., Sheets, Drive, Notion)
• Auth mechanisms change (e.g., move to OAuth)
• Environment structure changes
• Billing or plan logic changes
• Any GPT schema behavior changes