Export DeepSeek Harness sessions to Markdown for review, blogging and audit. dsh plugin.
Pending human review — the facts below come from the source repository.
Export DeepSeek Harness sessions to Markdown for review, blogging and audit. dsh plugin.
Users who want this capability in DSH; check the source and docs before installing.
dsh plugin --profile web add github:yangyongzhen/dsh-session-export No supported platforms were declared.
session/event 事件流，turn 结束并静默 debounceMs 后自动导出 - 输出 $DSH_HOME/exports/session-<id>.md（默认 ~/.dsh/exports/） - 进程退出（含 headless 一次性任务）时同步 flush 未导出的会话，不丢记录 - 渲染纯函数模块化导出，其他插件可复用 ## 安装 sh # 本地 checkout 安装（推荐，免构建） dsh plugin --profile <web|tui|headless> add file:/path/to/dsh-session-export # 或从 Git 仓库直接安装 dsh plugin --profile <web|tui|headless> add https://gitcode.com/qq8864/dsh-session-export.git ## 配置 通过 profile 的 cordis.patch.yml 或 --patch overlay 覆盖： yaml - id: session-export config: enabled: true # 总开关，false 关闭自动导出 outDir: ~/exports # 导出目录，默认 $DSH_HOME/exports debounceMs: 1000 # turn/end 后的静默等待（毫秒） maxResultChars: 2000 # 工具参数/结果的截断长度 includeInjected: false # true 时包含注入消息（上下文快照、skill 目录等） ## 导出内容 # 会话导出 - 会话 ID / 创建时间 / 工作目录 / Agent 预设 / 导出时间 - 模型（provider/model）与统计：turn/step/工具调用次数、token 消耗（含缓存读写）、耗时 ## Turn N / Step N.t 回合与步骤结构 ### 用户 / 目标 直接提示词与目标延续（注入消息默认过滤） #### 工具调用：name 工具名 + 参数 JSON #### 工具结果 返回内容（截断） ### 助手 模型文本 + 该步 token 明细 > Turn N 结束：完成 回合结束原因（完成/出错/被中断/达到上限…） ## 错误记录 全部回合错误汇总 ## 作为库使用 渲染管线独立于运行时，其他插件可直接复用： ts import { renderSessionMarkdown, summarizeStats } from 'dsh-session-export'; // 或按包路径导入：dsh-session-export 的 lib/markdown.js const md = renderSessionMarkdown(session.events, { sessionId: session.id }, 2000, false); ## 开发 sh pnpm install pnpm run build # tsc -> lib/ ## 验证 - dsh --profile <name> --dump-config 确认插件进入配置树 - headless 冒烟：dsh --profile headless "一句话介绍你自己" 后检查 $DSH_HOME/exports/Not by stars — picked by hand, one at a time.
A visual settings dashboard for DeepSeek Harness: every registered settings namespace — including third-party plugin ones the official UI never covers — rendered as editable, schema-driven forms. Early, but the right direction: stop hand-editing YAML.
A DIY trajectory visualizer for DeepSeek Harness: watch the agent’s plan and execution unfold as a graph instead of scrolling raw logs. Early, but it fills a real gap — understanding at a glance what the agent is doing.
Dispatch DSH work from Claude Code / Codex: spin up in-host DSH agent sessions with per-tier presets, watch native subagent progress, and borrow a multimodal bridge that lends the text-only harness vision and image generation. An orchestration plugin that turns DSH into a backend for other coding agents.