@michengai/dsh-pua
PUA 的 DSH 实现:完整协议、全局与会话配置、角色风味、子代理策略与独立验收 Loop
342 results
PUA 的 DSH 实现:完整协议、全局与会话配置、角色风味、子代理策略与独立验收 Loop
DSH multimodal input plugin: route file attachments (images, video, audio, text) through per-preset model chains and feed the results to the session model as prompt tokens. Adds a Multimodal settings page.
DeepSeek Harness (DSH) host-layer plugin: detects and auto-fixes Windows PowerShell command syntax mistakes — bash constructs, broken line continuations, pasted prompts, CRLF pollution — then executes the corrected command through the host shell seam under the normal sandbox/approval policy
Fail-closed network recovery, recommended answers, and safe unattended continuation for DeepSeek Harness.
Maestro Guard — pre-execute safety gate: native DSH approval prompts, parsed command-surface rules, decision journal
dsh plugin: token-efficient prompt-side shim for the official MCP client — folds mcp__* tool schemas out of every assembled prompt and exposes two constant meta-tools (mcp_list / mcp_call) instead
DSH 上下文管理插件:任务边界、项目帧、自动判别与上下文优化(开发中,R2 判别域 P8–P14b1 已施工)
一键把输入框草稿改写为更清晰、更贴 Agent 执行的提示词:直接替换 + 一键撤回,带保真闸与长度闸。
Give the DSH agent a clock: injects the current date/time into the runtime context on every turn. · 给 agent 一块表:每一轮把当前时间注入运行上下文。
DeepSeek Harness plugin: Set AI Language via system prompt injection
DSH Web GUI prompt-based conversation minimap: a vertical navigation rail on the left of long conversations, each anchor = one user prompt; hover shows the prompt preview, click jumps to that message.
Scheduled tasks for DeepSeek Harness: fire prompts into the agent at a time, an interval, or a daily clock time, with a web panel for tasks and run history.
全局提示注入 / 上下文病毒(context virus)防御插件:tools/pre-execute 扫工具参数、tools/post-execute 扫工具结果、agent/pre-step 扫进入模型前的消息并注入金丝雀守卫、llm/stream 出站消毒;支持 block / quarantine / monitor 三模式、磁盘化可演进病毒库(学习/导入/导出)与审计日志。纯 JS 无原生依赖,profile 层挂载,对所有预设与子代理全局生效。
Keeps long coding-agent sessions lean: wrap work in named tasks and, when one is done, fold its whole span into a short titled summary. The conversation stays readable, context costs stay low, and every fold's original content can be read back on demand. For DeepSeek Harness.
A safe external-skill discovery and adoption bundle for DeepSeek Harness
极简常用片段/命令工具箱 for DeepSeek Harness: JSONL-backed, five tools, no UI, no deps beyond the harness SDK. Save frequently used commands, code, URLs and prompts, then search them back in any session.
Conversation rollback for the DeepSeek Harness: rewind a session to the prompt before a chosen turn (restore the files that turn changed) and continue in a fresh forked session.
AI提示词优化器:诊断Prompt问题并输出结构化、高质量的优化版本
DeepSeek Harness 插件:Hermes 式跨会话持久记忆。MEMORY.md/USER.md 纯文本存储 + 冻结快照注入 system prompt + memory_* 工具(add/replace/remove/batch/show/refresh)
Per-directory context composition for DeepSeek Harness — .dsh/context.yaml declares systemPrompt + preload for every agent loop; memory is a baseline, not a preset
Append ambiguity-handling rules to dsh system prompt
Fridge-magnet sticker board for DeepSeek Harness web: one-click quick tasks docked above the composer. 输入框上方的冰箱贴纸栏:按长期/短期分类的快捷任务贴纸,点击即向当前会话发送预置 prompt 并执行。
Static and runtime security guard for the DeepSeek Harness: scans plugins and workspaces for malicious code, context injection and token waste (block/warn/clean), intercepts dangerous runtime tool calls and prompt steps, and exposes /scan, plugin_scan, the scan web panel and a user-managed allowlist. Fully static analysis — never executes scanned code.
Force DeepSeek Harness agents to think, reason, and reply in Chinese — a one-section dsh bundle