人格裂变工坊：用多人格裂变代替单次平均文案
人格裂变工坊是面向内容创作的 Agent Skill，由社区作者 user_19b9fbd9 发布维护，版本 1.0.0，分类 content-creation，无需 API Key。它针对单个 AI 产出的文案过于平均、难以覆盖多类受众动机的问题，用人格裂变流水线产出一稿多写。核心是五大心理测量系统生成的写作者人格卡（大五实证特质、九型恐惧欲望动机、MBTI 认知过程、DISC 可观察行为、霍兰德职业匹配，33 张卡 + 3 特邀席位），以及五阶段流水线：任务登记（先绑定成交/爆量/转化指标）、人格竞拍（受众动机匹配度低于 60 出局）、分裂写作（每张卡各产一版完整可用产物）、加权三维投票（最佳钩子/最佳转化/最佳风险预警）、主编综合（杂交最优版 + 条件式投放 + A/B 测试清单）。最终交付 md 产物集和 HTML 竞投投票报告。触发词包括人格裂变、裂变话术、文案裂变、一稿多写、多版本文案等。人格卡是写作透镜而非测评工具，不适用于事实查询、唯一标准答案任务或心理测量专业诊断。
Read Moredsh-plugin-compare: A Controlled A/B Comparison Tool for DeepSeek Harness Plugins and Presets
`dsh-plugin-compare` is the comparison plugin for DeepSeek Harness (DSH), used to validate plugin changes and preset differences. It supports evidence-based side-by-side timelines, A/B comparisons, and reports in multiple formats (HTML, JSON, etc.). Installation requires Node.js 22.19+, and it becomes available in the interface after starting via Web Host. A test environment must be prepared before use. Note that exported content may contain sensitive information and requires manual masking; statistical intervals do not automatically identify a winner; the system automatically copies the workspace and captures Git evidence, but ignores specific directories and symbolic links.
Read Moredsh-ab-sample: A/B Testing Sample Size Calculator
Developers using DeepSeek Harness (DSH) can use the `dsh-ab-sample` plugin to calculate A/B test sample sizes locally, without an internet connection and with precise results. The plugin is implemented purely in Node.js, provides the `ab_sample` tool, supports receiving test parameters, and returns detailed calculation results. Usage requires the dependencies `@deepseek-ai/cordis` (^4.0.1) and `@deepseek-ai/dsh-tools` (>=0.1.0-rc.6). The installation command is `dsh plugin add github:uckkk/dsh-ab-sample`, making it suitable for developers who need quick statistical calculations.
Read MoreDeepSeek Harness Plugin: dsh-plugin-abtest
dsh-plugin-abtest is a plugin for DeepSeek Harness (DSH) designed to address the challenge of validating plugin changes. It performs paired A/B experiments in isolated environments and supports evaluating task success rate, token consumption, latency, and tool error rate. It provides auditable evidence, deterministic release gates, and configurable rules to help teams determine whether a change is release-ready before publishing. It only generates offline recommendations and does not automatically modify systems. It requires a Node.js environment and is intended for DSH plugin developers.
Read More