---
title: "open-protein-binder-design"
description: "Replay, audit, and evaluate protein binder design campaigns with the openbinder CLI, source registry, protocol compiler, campaign replay, ranking audit, agent decision evaluation, external handoff, and evidence-graded fi"
canonical: https://agentpluginsdirectory.com/plugins/open-protein-binder-design
last-updated: 2026-10-08
---

# open-protein-binder-design
Replay, audit, and evaluate protein binder design campaigns with the openbinder CLI, source registry, protocol compiler, campaign replay, ranking audit, agent decision evaluation, external handoff, and evidence-graded final reports. No protein models, no GPU.
- Slug: open-protein-binder-design
- Publisher: Hylouis233
- Repository: https://github.com/MiniMax-AI/MiniMax-Code-Plugins
- Manifest: plugins/Hylouis233/open-protein-binder-design/plugin.json
- Version: 0.1.0
- License: MIT
- Category (editorial): other
- Skills: 7 (binder-agent-eval, binder-handoff, binder-protocol, binder-ranking-audit, binder-replay, binder-report, binder-source)
- MCP servers: 0
- Stars: 18
- Repository created: 2026-08-17
- Repository last pushed: 2026-09-21
- Publisher type: Organization
- Listing: https://agentpluginsdirectory.com/plugins/open-protein-binder-design
- Schema: https://agent-plugins.org/schemas/1.0.0/plugin.schema.json

## What open-protein-binder-design does, in the publisher's words

Replay, audit, and evaluate protein binder design campaigns in MiniMax Code, seven Skills on top of the openbinder CLI, covering source registration, protocol compilation, campaign replay, ranking audit, agent decision evaluation, external handoff, and evidence-graded final reports, without running any protein model.

Anthropic's public Claude binder design release (protocol, campaign records, candidates, precomputed scores, wet-lab labels) is a rare window into how an AI agent runs a real science campaign. But a raw data dump is not a workflow: from the files alone you cannot tell what was known and decided at each step, whether the published ranking survives recomputation, or how another model would decide given the exact same inputs. This Plugin turns the release into a replayable, auditable, comparable workflow, with strict gap discipline (missing is reported as missing, never interpolated) and byte-identical, wall-clock-free artifacts that diff cleanly under review.

From the project README, punctuation lightly normalized. Full text: https://raw.githubusercontent.com/MiniMax-AI/MiniMax-Code-Plugins/HEAD/plugins/Hylouis233/open-protein-binder-design/README.md

## Skills

- binder-agent-eval: 当需要在相同的阶段输入上评测 LLM agent 的结构化决策、或对比多个 provider 的决策质量时使用；对应 openbinder evaluate-agent（--provider/--compare/--case）与 openbinder provider doctor。典型触发语："用 mock provider 跑一遍 decision case 评测""对比一下几个 provider 在同一批 case 上的表现""先 doctor 一下 minimax 配置通不通"。同义场景：agent 评测、decision evaluation、visible/hidden 分层、泄…
- binder-handoff: 当需要把 binder 设计算力任务打包交给外部平台（Claude Science 或自有服务器）、或校验/导入外部返回的结果包时使用；对应 openbinder handoff export/validate/import。典型触发语："把这个 target 打包成 handoff 任务包""校验一下对方返回的 result package""把外部算回来的结果导入进来"。同义场景：external handoff、任务包导出、结果包校验、回传包导入、RUNBOOK。
- binder-protocol: 当需要把 binder 设计协议草稿编译成结构化 protocol.yaml、校验协议、或按阶段切出 stage slice 时使用；对应 openbinder protocol validate/compile/show 三个子命令。典型触发语："把这份 protocol 草稿编译一下""校验这个 protocol.yaml""把 cofold_scoring 阶段的切片拿出来"。同义场景：协议编译、protocol compiler、十阶段、stage slice、阶段上下文、protocol validate。
- binder-ranking-audit: 当需要重算并审计 binder campaign 的候选排序、核对 recorded 与 recomputed 是否一致时使用；对应 openbinder audit-ranking（--fixture/--source）。典型触发语："审计一下这个 campaign 的排序""recorded 和 recomputed 对得上吗""重算 top-N 命中率和 Spearman"。同义场景：ranking audit、排序审计、ensemble 重算、hit rate、缺失数据处理、missing data 报告。
- binder-replay: 当需要回放已记录的 binder 设计 campaign、重现每一步"当时知道什么、决定了什么"时使用；对应 openbinder replay（--fixture/--events/--source）。典型触发语："帮我回放一下 IL-7Ra 的 binder campaign""回放那个失败的 campaign""看看当时每一步都知道了什么"。同义场景：campaign replay、事件流回放、replay report、gap 分析、时间线重建。
- binder-report: 当需要把 output/ 下已产生的 replay、agent 评测、ranking 审计产物汇总成最终报告束时使用；收集既有 artifact、按六类证据分级标注、生成 limitations 并提交 stage-gate 审批。典型触发语："把这些产物汇总成最终报告""出一份 binder 最终报告和 limitations""把 replay/评测/审计结果打包成报告束"。同义场景：最终报告、报告束、evidence grading、limitations、claim check、收尾报告。
- binder-source: 当需要登记、查看、锁定或按白名单下载 binder 数据源文件时使用；对应 openbinder source list/inspect/lock/download 四个子命令。典型触发语："看看 anthropic-claude-binder-v1 这个源""inspect 一下远端文件列表""把源锁定到当前 revision""下载那个 README 小文件"。同义场景：数据源注册、source registry、pin revision、source.lock.json、下载白名单、download policy、blocked 文件、源不可用怎么办。

Descriptions come from the frontmatter of each SKILL.md, punctuation lightly normalized.