@dsh-external/dsh-plannotator
Unofficial Plannotator integration for DeepSeek Harness: annotate plans inline and return structured feedback to the waiting coding agent
318 results
Unofficial Plannotator integration for DeepSeek Harness: annotate plans inline and return structured feedback to the waiting coding agent
更好的折叠栏:将工具调用与 Think 段自动折叠为步骤折叠栏,回合结束后整回合收拢为带实时指标(耗时/首字/token/tok/s/缓存命中)的回合折叠栏,只保留最终总结正文;前导图标可自定义,内置动态扑克牌(运行中花色循环卡牌动画、收起牌堆、展开扇形),提供官方折叠方式设置行,适配 DSH 新旧版本。零侵入 DSH 源码。
DSH Web 会话时间线整理:已完成轮次自动折叠(只留最终结论)+ 过程/结论分隔线 + 智能加载更早历史;settings 与折叠 DOM 自动适配 DSH 0.1.0-rc.7 ~ 0.1.2-rc.1。左缘定位条在旧版 DSH(无官方右缘 TurnNavigator)可用,DSH 0.1.2+ 暂停。
DeepSeek Harness plugin: per-event task-completion and attention sounds for turn-end / approval / question / plan-review / goal-blocked / task-failure, each with its own sound (built-in synth, local audio file) and volume (Web UI)
Search, favorites, branching, and personalization for the official DeepSeek Harness turn navigator
DeepSeek Harness-native publicity toolkit: turn any long-form source into a source-grounded, platform-native promotion matrix.
DSH 微信桥插件:turn/approval 推送到微信,微信远程监控/批准/驱动 agent(iLink 通道,持久去重/分段/合并/白名单)
Self-evolving memory + skill lifecycle for DeepSeek Harness. Cross-session memory with zero-token deterministic recall (bigram-Jaccard fused with FTS5 BM25 via RRF), a tiered approval gate, and reinforcement that strengthens what you repeat. Procedural knowledge crystallizes into SKILL.md files that refine in place and are curated through an active-stale-archived lifecycle (reversible archive, pre-op backups, rollback, never deletes). Includes background per-turn review, anti-bloat convergence for both skills and memory, an auto-grown user profile, and a web settings page.
Self-contained Trellis workflow trigger for DeepSeek Harness: per-turn breadcrumb injection via agent/pre-step, a bundled skill provider, and DSH-native task/phase state resolution. No Trellis AGPL code is vendored; workflow semantics are rewritten and licensed MIT.
Inline +N −M diff badges on edit/write tool rows and a per-turn file change summary card for DeepSeek Harness. PTC dispatch fallback (code-dispatch/ptc-dispatch), no git required.
Codex-style per-turn code change summaries, review tab, and guarded undo for DeepSeek Harness
DeepSeek Harness plugin that turns a 3D idea into a shootable scene brief and publish-ready showcase copy.
OpenTelemetry and Langfuse observability exporter for DeepSeek Harness: turn/step/tool/LLM spans, token and cost metrics, and sanitized LLM prompt/completion capture from the session/event stream, with async batching, bounded offline buffering, and retry with backoff.
Piano-key style in-conversation navigation bar for the DeepSeek Harness web GUI: one key per user message, hover ladder animation, turn-preview tooltip, active-message highlight, click-to-jump. Official dual-face dsh plugin (host + client), no dsh source changes.
Cosplay mode for DeepSeek Harness: a global switch turns the agent into any character you define, with a settings UI for role configuration and management.
Enhanced ACP server for DeepSeek Harness: block-level streaming, usage/stat telemetry (cache hit rate, token speed, input/output tokens, context length, turns, tool timing), model & reasoning-effort switching, and permission-preset control over the ACP wire (Zed-friendly)
Writing skill for DeepSeek Harness: de-AI humanizer plus personal voice clone. Scans AI-writing patterns, builds a style fingerprint from your samples, and returns rewrite briefs so the agent writes like you.
Task-completion reminder for DeepSeek Harness: in-page toast, OS notification, and sound when an agent turn or background job finishes
DSH web plugin: roll back to before any of your questions, or edit the last one and resubmit, restoring the workspace files the discarded turns produced.
View and manage archived sessions in DSH: browse the archive, read full transcripts in a chat-style dialog (turns, reasoning, tool executions), and remove unwanted records. Works on desktop and web.
Personal Harness workbench: hero banner, completed-turn confetti, and a built-in task board
Privacy-minimal heuristic per-turn execution summaries for DeepSeek Harness
Hatch a digital creature in your DeepSeek Harness. It feeds on your real work — turns, tool calls, even your errors — and evolves along branching lines shaped by how you actually work. Zero tokens, zero tools, the model never knows.
DSH web plugin: edit the last user message back into the composer after stopping a turn