Plugins

Survival mode: all sessions share one satiety/health pool — each model step burns satiety, starvation freezes tool calls until fed, with an MC mini-game for gathering food.

Folds streaming reasoning and tool calls into collapsed step bars and collapses each finished turn into a single summary bar; step bars show native tool icons, filenames, edited line counts, and failure counts, finished turns restore duration, first-token time, token usage, generation speed, and cache hit metrics from recorded Session events, and preferences persist on the Host. Changes browser presentation only. Requires DSH 0.1.5-alpha.1.

Generates a standard motor-repair procedure from the diagnosis and motor rating, with labour time, staffing and a quality-check point for each step.

Coding retrospect distilled memory for DSH: distills rejected solutions, pitfalls and constraints from each turn via a local LLM, tags them into MemOS (type:rejected_solution), and boosts weighted recall on coding tasks via agent/pre-step.

Actionable quality diagnostics for DSH agent sessions, with tool-call evidence and next-step recommendations.

Composer code-quote collapse for the DSH Web UI: pasting a `path:line` header followed by code folds the input into one compact inline quote chip, and at send time the `agent/pre-step` hook injects the full pasted snapshot to the model as a separate context message. 0.5.x also folds sidebar file-selections that insert only a `path:lines` line (large selections), fetching the code from disk by session cwd, and preempts Ctrl+V paste interception.

Adds a sidebar tab that folds a session into user turns (input, tool steps, output) with one-click jump back into the conversation.

Force the agent's chain-of-thought, planning, tool-invocation reasoning, and internal deliberation to stay in Chinese — the language constraint is injected at the top of the system prompt before every reasoning step, so it cannot be diluted by tool-returned content. A settings panel lets you edit the constraint text on the fly.

Adds parametric CAD modeling, geometry validation, previews, design history, and STEP/STL export to DSH.

Portrait-adaptive communication skill: builds a live user portrait (proficiency, per-domain gaps, current state) from the conversation and calibrates every answer's depth, jargon density and step granularity to it; behavior-over-claims dumb-play detection; transparent profile snapshot export. MIT.

Session manager panel under Settings → Plugins: list all sessions with title, cwd, agent preset, updated time, turns and steps, expand a row to view message history, delete non-running sessions, and export a session as JSON or Markdown via a browser Blob URL.

Feishu/Lark channel that drives DeepSeek Harness agents from chats: streaming rich cards with live thinking and tool steps, per-turn cost metering at DeepSeek peak/off-peak rates, one-tap approval cards, model-gated image intake, and QR onboarding without a public endpoint.

Step-0 observation probe for dsh-compaction-audit: logs every compaction/* session event and probes whether shadowedSeqs still resolve to readable events in session.events. Observe-only; never injects, never blocks.

Publishes a project plan the agent writes: ordered steps in a panel docked to the right of the frame, and the framework graph as an auto-laid-out diagram in its own centre tab, downloadable as SVG or PNG. Charts are rebuilt from the session log, so they survive restart and resume.

Adds read-only, scriptable playback to native DSH conversations, with simulated user typing, forward and reverse playback, event stepping, speed control, idle-gap compression, and direct seeking by event, turn, or recorded time.

Metadata-first runtime profiler for DSH agents: turn/step traces, model and tool wall-time timelines, provider usage, retries, errors, and live or historical sessions.

Auto-supplement QMD semantic search when the model greps/globs a knowledge-base directory: zero-dependency DSH plugin, async injection via next-step, no manual invocation.

Long-term memory injected at session start: preferences/facts/summaries/knowledge in global and per-project scopes, budgeted recall via `agent/pre-step`, durable JSON store.

GPU-aware execution layer for DeepSeek Harness: gpu_status / gpu_exec / gpu_run_bg tools, per-step GPU context injection, automatic CUDA_VISIBLE_DEVICES card selection

Telemetry backend exporting turns, model steps, and tool calls to yiTrace.

DeepSeek Harness (DSH) plugin: an Overview tab beside Chat/Trajectory plus a right-sidebar page summarizing session progress (goal, todos, key steps), artifacts, and references, with per-turn scoping, OpenViking memory preview, and quick-nav linkage.

DSH plugin: delete an archived session's local files — transcript, projection cache, and its subagent sessions — after a two-step confirmation.

CAD 3D model preview for dsh-better-sidebar: renders STEP/STP/IGES/BREP (OpenCascade WASM) and STL/3MF/OBJ/PLY with an orbit camera in the DSH right sidebar.

Task-list staleness guard for DeepSeek Harness: when the model stops updating its todo list but keeps working, the guard appends a short reminder before the next step so the list gets written back to the real progress. / 任务清单陈旧度守卫：模型建了清单却忘记回写时，在下一步之前自动追加一