agent-ultramode 0.1.0 → 0.1.2

package/README.md

@@ -1,5 +1,9 @@
 # ultra
 
+[![npm](https://img.shields.io/npm/v/agent-ultramode?logo=npm&color=cb3837)](https://www.npmjs.com/package/agent-ultramode)
+[![release](https://img.shields.io/github/v/release/maverick-tr/agent-ultramode?color=blue)](https://github.com/maverick-tr/agent-ultramode/releases)
+[![license](https://img.shields.io/badge/license-MIT-blue)](./LICENSE)
+
 Took the [LLM-as-a-Verifier](https://github.com/llm-as-a-verifier/llm-as-a-verifier) paper (Kwok et al., 2026) and turned it into a simple `/ultra` command for coding agents.
 
 Instead of shipping the agent's first attempt, `ultra` runs your task **N times in isolated git worktrees**, in parallel, then uses the **same model** as a verifier to pick the best result. When it is confident (or when most attempts succeeded) it applies the winning diff to your working tree; when the attempts diverge and it is not sure, it hands you the top candidates. No cross-model dependency, no extra services, no first-attempt lottery.
@@ -8,13 +12,13 @@ And unlike most "add a verifier" posts, I benchmarked it before believing it. Th
 
 ## Demo
 
-<video src="https://github.com/maverick-tr/agent-ultramode/raw/main/ultramode.mp4" controls muted playsinline width="100%"></video>
+<video src="https://github.com/user-attachments/assets/134487dc-1d60-434b-a2dc-97cfd5942959" controls muted playsinline width="100%"></video>
 
-> For a reliable inline player on GitHub, edit this README on github.com and drag `ultramode.mp4` into the editor; GitHub inserts a `https://github.com/.../assets/...` URL that plays inline. The file is committed at [`ultramode.mp4`](./ultramode.mp4).
+[▶ Watch the demo](https://github.com/user-attachments/assets/134487dc-1d60-434b-a2dc-97cfd5942959) (55s, opencode + DeepSeek V4 Flash)
 
 ## The receipts
 
-Benchmarked on **Terminal-Bench** with **DeepSeek V4 Flash 0731** as both the agent and the verifier, running inside **opencode** (N=5 attempts, same-model verifier). Verified end to end with both **OpenCode** and **Claude Code** as the agent (the numbers above are the OpenCode run; a Claude Code number will be added once benchmarked). Works with any agent, not just opencode.
+Benchmarked on **Terminal-Bench** with **DeepSeek V4 Flash 0731** as both the agent and the verifier, running inside **opencode** (N=5 attempts, same-model verifier). Verified end to end with **OpenCode**, **Claude Code**, and **cline** as the agent (the numbers above are the OpenCode run; a Claude Code number will be added once benchmarked). Works with any agent, not just opencode.
 
 | Slice | base@1 (single shot) | ultra (best-of-N + verify) | oracle@5 (ceiling) |
 |---|:---:|:---:|:---:|
@@ -161,7 +165,7 @@ The 15 tasks were deliberately failure-skewed, so the +9 points is not what you
 
 - [x] **Standalone CLI** (`npx agent-ultramode`) so the loop runs anywhere, with any agent, no opencode required.
 - [x] **Multiple models in one pass** (repeatable `--agent`): spread attempts across different models, one neutral verifier picks the best.
-- [x] **OpenCode and Claude Code** verified end to end.
+- [x] **OpenCode, Claude Code, and cline** verified end to end.
 - [ ] **First-class agent integrations** with tuned defaults, and a benchmark number, for Claude Code, Grok, Pi, and Codex.
 - [ ] Native slash-command or MCP packaging per agent. Contributions welcome.
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "agent-ultramode",
-  "version": "0.1.0",
+  "version": "0.1.2",
   "description": "Best-of-N for coding agents with a same-model verifier: run your task N times in isolated git worktrees and apply the verified winner. A /ultra command for opencode.",
   "module": "ultra.ts",
   "main": "ultra.ts",