/plan-synthesize — best-of-all-worlds

Feed competing plans into a single GPT Pro session and produce a hybrid that integrates the strongest ideas from every input.

Workflow

1. Pick a primary. Choose the best of the competing plans as your "base" — usually GPT Pro's. Its content gets the git-diff revisions applied to it.

2. Open the synthesis prompt. Read .flywheel/prompts/plan-synthesize-best-of-all-worlds.md. Copy it into a fresh GPT Pro (web) conversation.

3. Paste the base plan, then each competitor. Label each clearly.

4. Receive git-diff revisions. GPT Pro will produce a list of diff-style changes integrating the best ideas.

5. Apply in-place with Claude / Codex. Use the prompt in .flywheel/prompts/plan-integrate-synthesis.md. The agent applies the diffs to .flywheel/plans/<date>-<slug>-gpt-pro.md (renaming it to -hybrid.md is a good practice) and reports which revisions it wholeheartedly agrees with, somewhat agrees with, or disagrees with.

6. Iterate. Use .flywheel/prompts/plan-refine.md in fresh GPT Pro conversations 4–5 more times until convergence.

7. Sanity overshoot. Once per synthesis, run .flywheel/prompts/overshoot-mismatch-hunt.md to catch elements the model silently dropped.

Output

A single hybrid plan at .flywheel/plans/<date>-<slug>-hybrid.md. This is the one you'll feed into /plan-to-beads.