106 results
DSH web plugin: an Archived Sessions manager in Settings — browse archived conversations, expand per-session details (disk usage, downloads, tool usage, lineage, cross-session recall), multi-select and permanently delete.
Model-driven context-window rollover bundle for the DeepSeek Harness: fresh working context, durable model-managed notes, a small verbatim recent tail, and targeted history recovery — no summarization.
Review agent-produced file changes (line-level red/green diff + undo) both as a chat turn-tail row and as a dsh-better-sidebar tab; a port of left0ver's dsh-file-review.
Private, zero-trust gateway for DeepSeek Harness Web: loopback-only proxy, exact principal allowlists, and provider-agnostic onboarding
DeepSeek Harness 远程与移动端安全网关插件：零修改 DSH 底层代码安全开放局域网与 Tailscale 连接 | DeepSeek Harness (DSH) Remote & Mobile Security Guard: safely opens Tailscale/LAN with zero core modifications, QR scan auth, RSA encryption & brute-force defense.
Mobile adaptation for the DeepSeek Harness Web GUI: single-column layout, sidebar drawer, compact header/composer, fullscreen settings, trajectory floating details, stats pill panel
Ponytail, lazy senior dev mode, for DeepSeek Harness: 6 skills (ponytail, ponytail-audit, ponytail-debt, ponytail-gain, ponytail-help, ponytail-review) adapted from DietrichGebert/ponytail (MIT)
Archived-sessions manager for the DeepSeek Harness web UI: list / unarchive / delete / detail from Settings. · DSH 归档会话管理：设置页查看、释放、删除已归档会话。
DSH 完整移植版 dietrichgebert/ponytail — 懒惰 senior 模式，6 个中文 Skill，无空 tool
Rewind a DSH conversation to any user message: hide the tail from the chat view and the model context, keep the durable log intact, cancel freely before sending.
Mobile plugin for the DeepSeek Harness Web shell: off-canvas sidebar/details drawers, bottom nav bar, and touch-first responsive sheet
Tool search & slimming for DeepSeek Harness: Hermes-style progressive disclosure — search, describe, and call long-tail tools on demand while keeping core tools eager
DeepSeek Harness 编程模式组合包：标准模式之上强制执行 Superpowers 工程纪律（TDD、系统化调试、先计划后编码、完成前验证、代码审查），常驻 Ponytail 代码量纪律，内置全套 superpowers 与 ponytail 技能，dsh plugin add 一键安装。
Runtime compatibility shim for dsh (DeepSeek Harness) cordis bundle plugins: decouples third-party plugins from real dsh internal service names, module paths, and RPC details via a version-aware adapter registry.
Keep long answers and detailed prompts side by side with a dockable DSH Composer.
科研收敛驾驶舱 (developer preview): hard-gated research convergence loop for DSH agents - 7 conv_* tools + host session projection + turn-tail card (conversation.chat.turnTail) + cockpit strip (conversation.input.dock) + settings section. 零 npm 依赖 (node builtins
Widescreen mode for the DeepSeek Harness session detail page: a General-settings toggle that lets the chat use the full browser width instead of the fixed 748px content column.
Auto-inject user-chosen skills (e.g. caveman, ponytail) into DeepSeek Harness sessions: every prompt or once at session start, with a settings page and a composer indicator. 在 DSH 会话中自动注入所选技能（如 caveman、ponytail）：每轮提示或仅在会话开始时，含设置页与输入区指示。
Agent-driven long-term memory for DeepSeek Harness: scoped memory (global + per-workspace), layered entries (fact/knowledge/episodic), time-bucket compaction (day→week→month→year), associative recall (related chains) + memory_relate navigation (multi-hop BFS closure), auto recall injection on user messages (CJK bigram search, tail append), agent-decided content.
A safe, accessible session manager for DeepSeek Harness with archive controls, activity details, lineage, and output-file cleanup.
General APEX preset with a Minimal Pro anchor, role-bounded PTC implementation, verified handoffs, and evidence convergence
Background stream listeners that wake the DeepSeek Harness agent — put a watch on a command or file and receive new lines as in-session notices. Ships a daemon host so a watcher can run unattended for weeks.
MiniMax token-plan quota dock for DSH web — drop-shape 5h usage %, click for detail panel, auto-hides for non-MiniMax models.
Curate tool output before it enters the model's context: keep error/warning lines and head/tail, spill the full text to disk with a pointer. Faster turns, smaller context, sharper attention — for every tool, not just bash.