[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-85159":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":10,"languages":9,"totalLinesOfCode":9,"stars":11,"forks":12,"watchers":13,"openIssues":14,"contributorsCount":15,"subscribersCount":15,"size":15,"stars1d":15,"stars7d":15,"stars30d":15,"stars90d":15,"forks30d":15,"starsTrendScore":15,"compositeScore":16,"rankGlobal":9,"rankLanguage":9,"license":17,"archived":18,"fork":18,"defaultBranch":19,"hasWiki":20,"hasPages":18,"topics":21,"createdAt":9,"pushedAt":9,"updatedAt":22,"readmeContent":23,"aiSummary":9,"trendingCount":15,"starSnapshotCount":15,"syncStatus":24,"lastSyncTime":25,"discoverSource":26},85159,"fusion-fable","duolahypercho\u002Ffusion-fable","duolahypercho","Fuse two frontier models into one Fable-tier answer: Opus 4.8 drafts, a second model (Opus 4.8 or GPT-5.5 via codex) checks, Opus fuses. A Claude Code skill.",null,"Shell",245,35,152,1,0,40.34,"MIT License",false,"main",true,[],"2026-06-15 10:05:02","# Fusion-Fable\n\n**Fuse a panel of frontier models into one Fable-tier answer.**\n\nFusion-Fable is a [Claude Code](https:\u002F\u002Fclaude.com\u002Fclaude-code) skill that runs a hard question through a\n**panel → judge** pipeline. The same prompt is dispatched to several models *in parallel* — each answering\nindependently with web search and bash, none seeing the others' work — and then Opus 4.8 judges every\nanswer into a structured analysis (consensus, contradictions, partial coverage, unique insights, blind\nspots) and writes a final answer grounded in it.\n\nThe mechanism is **independence, then synthesis**. The diversity that makes a panel beat a single model is\nharvested, not manufactured: running the same prompt independently yields different reasoning paths, tool\ncalls, and sources — even two cold runs of the *same* model diverge enough that synthesizing them beats\nrunning it once. So there are no contrived \"lenses\" or personas; every panelist gets the task verbatim and\nanswers it straight. Fuse **Opus 4.8 + Opus 4.8**, or **Opus 4.8 + GPT-5.5** (via the `codex` CLI), into a\nresult better than either alone — a Fable-tier fusion.\n\n```\n                      ┌──────────────┐\n                 ┌──▶ │  panelist 1  │ ─┐   (web + bash, independent)\n                 │    └──────────────┘  │\n                 │    ┌──────────────┐  │   ┌──────────────┐\n prompt ──▶ fan ─┼──▶ │  panelist 2  │ ─┼─▶ │   Opus 4.8   │ ──▶ final answer\n            out  │    └──────────────┘  │   │   (judge +   │     (grounded in\n                 │    ┌──────────────┐  │   │  synthesize) │      the analysis)\n                 └──▶ │  panelist 3  │ ─┘   └──────────────┘\n                      └──────────────┘\n              Opus 4.8 \u002F GPT-5.5 \u002F Gemini      consensus · contradictions ·\n              (each answers blind)             partial · unique · blind spots\n```\n\nOpus 4.8 **always** judges and writes the final answer — the pipeline can't be reversed, because the\npanelist models can't call back out to spawn Opus.\n\n## The panels\n\n| Slug | Panel | Requires |\n| --- | --- | --- |\n| `opus4.8-4.8` | the **same prompt run twice** as 2 independent Opus 4.8 panelists → Opus judges | nothing — works everywhere |\n| `opus4.8-gpt5.5` | Opus 4.8 + **GPT-5.5** (codex) in parallel → Opus judges | the `codex` CLI |\n| `opus4.8-gpt5.5-gemini3.1pro` | Opus 4.8 + GPT-5.5 + **Gemini 3.1 Pro** in parallel → Opus judges | `codex` + `gemini` CLIs |\n\nThe skill auto-detects which panelist CLIs are installed and uses the richest panel available, falling\nback gracefully when one is missing.\n\n## Install\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fduolahypercho\u002Ffusion-fable.git\ncd fusion-fable\n.\u002Finstall.sh\n```\n\nThis copies the skill to `~\u002F.claude\u002Fskills\u002Ffusion` and the slash commands to `~\u002F.claude\u002Fcommands`,\nthen prints which panels your machine can run. Restart Claude Code (or run `\u002Freload-skills`) afterward.\n\n> Override the target with `CLAUDE_CONFIG_DIR=\u002Fpath\u002Fto\u002F.claude .\u002Finstall.sh`.\n\n## Use it\n\nThree ways, all equivalent under the hood:\n\n- **Natural language** — just ask. The skill auto-triggers and picks the richest panel:\n  > \"Run this through Fusion: is it safe to `ALTER TABLE … ADD COLUMN` on a 200M-row Postgres table in prod?\"\n- **Pinned slash commands:**\n  ```\n  \u002Ffusion-opus4.8  does my JWT refresh-rotation design have a replay hole?\n  \u002Ffusion-gpt5.5   is git push --force-with-lease actually safe on a shared branch?\n  ```\n- **Force a panel in prose** — \"run the `opus4.8-gpt5.5` Fusion on …\".\n\nEvery run returns the same structure: a **Final answer** up top, then the audit trail —\n**Consensus \u002F Contradictions \u002F Partial coverage \u002F Unique insights \u002F Blind spots** — with each point\nattributed to the panelist that raised it, so you can see how the answer was assembled.\n\n## Requirements\n\n- **Claude Code**, with the session running **Opus 4.8** (panelist subagents and the judge inherit the\n  session model — on another model the slug is nominal, not literal).\n- For `opus4.8-gpt5.5`: the [`codex` CLI](https:\u002F\u002Fgithub.com\u002Fopenai\u002Fcodex) installed and logged in to an\n  account with GPT-5.5 access. The runner uses `codex exec` (tested against `codex-cli` 0.139).\n- For the 3-model panel: a `gemini` CLI installed and authenticated. Adjust the invocation in\n  `skills\u002Ffusion\u002Fscripts\u002Frun_gemini.sh` to match your CLI's flags.\n\nOnly the **`opus4.8-4.8`** panel is truly zero-setup; the GPT-5.5 and Gemini panels light up once their\nCLIs are installed and authenticated.\n\n## What's in here\n\n```\nskills\u002Ffusion\u002F\n  SKILL.md                  fan out in parallel → judge → grounded final answer\n  scripts\u002F\n    detect_panel.sh         picks the richest available panel\n    run_codex.sh            runs the GPT-5.5 panelist (web + bash), captures its answer\n    run_gemini.sh           runs the Gemini panelist (graceful no-op until the CLI exists)\n  references\u002F\n    panel.md                why independent parallel runs (no lenses) — the panel mechanism\n    judge_rubric.md         the structured analysis + grounded final answer\ncommands\u002F\n  fusion-opus4.8.md         \u002Ffusion-opus4.8  (pinned opus4.8-4.8 panel)\n  fusion-gpt5.5.md          \u002Ffusion-gpt5.5   (pinned opus4.8-gpt5.5 panel)\ninstall.sh                  copies the above into ~\u002F.claude\n```\n\n## Why a panel beats one model\n\nOn the DRACO deep-research benchmark, OpenRouter found that fusing model answers consistently beats the\nindividual models — and that a meaningful chunk of the lift comes from the *synthesis step itself*, not\njust from mixing architectures: two independent runs of one model, synthesized, beat that model run once.\nFusion-Fable implements that same independence-then-judge pipeline locally in Claude Code.\n\n## Cost & latency\n\nA panel costs roughly N× a single answer in tokens and runs as slow as its slowest panelist. That's the\ndeliberate trade: spend more to stop being confidently wrong where that's expensive. For quick or\nlow-stakes questions, a single direct answer is the right call.\n\n## License\n\nMIT — see [LICENSE](LICENSE).\n",2,"2026-06-15 02:30:12","CREATED_QUERY"]