Consult Oracle

Consult ChatGPT Pro via ChatGPT browser automation for problems that resist standard approaches.

Configuration

The oracle reads from ~/.turbo/config.json:

{
  "oracle": {
    "chatgptUrl": "https://chatgpt.com/",
    "chromeProfile": "Default"
  }
}

Step 1: Refresh Cookies

Before running the oracle, refresh ChatGPT cookies from Chrome. Requires timeout of 60000ms. A macOS Keychain password prompt will appear for the user.

python3 scripts/refresh_cookies.py

If the script reports session validation failure, the user needs to log into ChatGPT in Chrome first.

Step 2: Identify Key Files

Find the 2-5 files most relevant to the problem.

Step 3: Formulate the Question

Write a clear, specific problem description. Include what has already been tried and why it failed. Open with a short project briefing (stack, services, build steps). The more context, the better the response.

Step 4: Run the Oracle

Use a generous timeout (60 minutes / 3600000ms). The script loads chatgptUrl from ~/.turbo/config.json automatically. Generate a random tag and persist the response:

ORACLE_TAG=$(head -c 4 /dev/urandom | xxd -p) && mkdir -p .turbo/oracle
python3 scripts/run_oracle.py --prompt "<problem description>" --file <relevant files...> --write-output ".turbo/oracle/$ORACLE_TAG.txt"

Step 5: Run /evaluate-findings Skill

Read the response from .turbo/oracle/$ORACLE_TAG.txt. Run the /evaluate-findings skill on it. Oracle suggestions are starting points — cross-reference with official docs and peer open-source implementations before accepting.

Step 6: Run /apply-findings Skill

Run the /apply-findings skill on the evaluated results.