القائمة
Synthesize competing plans from multiple frontier models into a single best-of-all-worlds hybrid. Use after /plan has produced 2+ competing plans in .flywheel/plans/.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
| name | plan-synthesize |
| description | Synthesize competing plans from multiple frontier models into a single best-of-all-worlds hybrid. Use after /plan has produced 2+ competing plans in .flywheel/plans/. |
Feed competing plans into a single GPT Pro session and produce a hybrid that integrates the strongest ideas from every input.
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.
Open the synthesis prompt. Read
.flywheel/prompts/plan-synthesize-best-of-all-worlds.md. Copy it
into a fresh GPT Pro (web) conversation.
Paste the base plan, then each competitor. Label each clearly.
Receive git-diff revisions. GPT Pro will produce a list of diff-style changes integrating the best ideas.
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.
Iterate. Use .flywheel/prompts/plan-refine.md in fresh GPT
Pro conversations 4–5 more times until convergence.
Sanity overshoot. Once per synthesis, run
.flywheel/prompts/overshoot-mismatch-hunt.md to catch elements
the model silently dropped.
A single hybrid plan at
.flywheel/plans/<date>-<slug>-hybrid.md. This is the one you'll
feed into /plan-to-beads.
Deep review — random exploration + cross-agent review, alternating. Finds integration bugs that self-review misses. Run every 30–60 min during active implementation.
Overview of Forge's Agentic Coding Flywheel methodology. Use to route between plan, bead, and code space. See FLYWHEEL.md at the repo root.
End-of-session ritual. Close finished beads, file new ones for remaining work, run verify, commit in logical groupings, push. Work is not complete until git push succeeds.
Start a new Flywheel planning round. Produce initial markdown plan, route to competing frontier models, synthesize hybrid, refine 4–5×. Use before any coding begins on a new feature.
Translate a polished markdown plan into self-contained beads in .beads/beads.jsonl. Every bead must be executable without reopening the plan. Use after the plan has converged.
Polishing pass over all open beads. Run 4–6 times. "Check your beads N times, implement once." Use after plan-to-beads and between refinement rounds.
