All plugins

Export DSH conversations as Smartisan Notes-style PNGs, or create and update Markdown notes in a configured account-scoped workspace.

Talk to Excel in DeepSeek Harness: create, edit, repair, and verify spreadsheets by conversation, with automatic formula health checks after every edit.

In-conversation generative UI: the model renders interactive HTML cards into the chat stream, with streaming preview and sandboxed rendering.

Side panel with file browser, terminal, and Git review for quick file previews.

A "focus chat" minimal view that shows only final outputs.

Interactive UI components rendered inline in replies: layout, charts, forms, quizzes, mermaid, 3D scenes, and an action event loop back to the model.

Select text → annotate → send with your message; replies map back to each annotation.

Conversation node navigation bar for quick jumps between user messages.

Background task status bar: progress plus live output tail on the chat page.

Collapse noisy messages (Think, Bash, etc.) in conversations.

Keyboard-first command palette for the DSH Web UI.

Document reading mode for DSH.

Cross-platform file drag-and-drop with raw path insertion, no file copying.

PiUI-style diff viewer replacing the stock DiffBlock for write/edit tool calls.

Codex-style `@file` mentions: search workspace files in the composer and attach their contents to prompts.

Right-side dot-timeline rail: jump between user messages.

DeepSeek API balance, balance trend, and daily usage charts in DSH Web settings.

DeepSeek API balance and 24-hour usage dashboard in Settings, with estimated spend, token counts, request counts, and an hourly timeline.

Claude Code-style full-screen terminal UI: pixel-whale header, live status line, and streaming thought expansion.

Full sidebar workbench with file rendering and editing, terminal, Git, and subagents; third-party plugins can register new tabs.

One-click collapse of every expanded section (Think rows, tool cards) with a live-count pill and a customizable hotkey.

Rewind conversation and workspace state, powered by a persistent Change Ledger.

Cross-session messaging for DSH: any session on the machine can list and message any other, Claude Code-style, via a local heartbeat registry and inbox.

Automatic conversation distillation: background subagent reflection + skill create/update.

Share your conversations with one click.

Branch-based message editing, reroll, retry, and a version timeline.

Deep Mnemon integration: local three-tier memory (Runtime Memory, retrievable Documents, supervised Memory Spaces).

Cross-session memory: SQLite with a human-editable Markdown mirror, background consolidation (dedup, merge, conflict resolution), and six memory tools.

Turn navigation for the DSH Web UI.

`/side` persistent side sessions and `/btw` one-shot side questions, run in a temporary fork without touching main history.

Share any excerpt of a conversation.

Local-first learning mode: cross-session learning threads with per-source explanations.

Import Claude Code / Codex / ChatGPT / Cursor / Gemini / Reasonix / opencode chat histories as resumable DeepSeek Harness sessions.

Cross-instance message and event handoff between DSH instances via an interconnect server.

One shell tool for PowerShell / Git Bash / WSL on Windows plus an interactive PTY terminal; the default terminal is chosen by the user in DSH settings.

Vision tasks for text-only models: intent-aware image Q&A, long-screenshot OCR, UI reproduction, grounding, and pixel diff.

Create and manage sandboxed JavaScript tools with a Monaco editor and model-driven tool lifecycle.

Accessibility-first macOS computer use: fresh observations, stale-state rejection, scoped permissions, and safe input.

Let the AI connect to databases and write SQL for you.

Parse/query/aggregate/convert CSV (RFC 4180) with a zero-dependency state-machine parser.

Safe math expression evaluator, zero-dependency recursive-descent parser.

Cited memory over DSH's lossless session log: distilled facts carry `(sessionId, eventRange)` citations that expand back to the exact original log excerpt.

JSON queries with a JMESPath subset.

JSON Schema validation: validate/paths/explain/normalize.

Descriptive statistics, percentiles, frequency distributions, and correlation.

Strict ISO 8601 parsing, IANA timezone conversion, and UTC calendar arithmetic.

Build auditable KB packs (SQLite FTS5) from md/txt/docx/pdf with deterministic retrieval and original-text reading.

Expose MineRU document parsing tools to the model.

Bounded, cell-addressed `doc_read`/`doc_write` for xlsx / pdf / docx / pptx / ipynb, plus an MCP server and CLI.

Per-agent on-demand tool discovery and progressive schema disclosure.

OpenMAIC: classrooms, slides, interactive widgets, and Socratic teaching.

Academic assistant plugin.

Continual self-evolution: versioned, auditable, rollback-safe harness state (prompts, memory, skills, subagent specs) refined from session trajectories, with review gates and hot-reloaded skills.

base64/url/hex encoding, common hashes, and UUID generation.

Action-state time memory: record typed states and actions, then analyze trends, anomalies, and causality.

Test/extract/safe-replace/statically explain regexes without executing code.

Per-call model, provider, persona, and toolFilter overrides for subagent delegation, with @preset: references and provider/model composite ids.

Extends dsh-subagent-tools with a per-call cwd for subagents, shipped with the two in-process provider patches it requires.

Voice notes in, spoken answers out: dictate audio that becomes user messages (transcribe), have the agent read replies aloud (speak), local-first under ~/.dsh/voice.

Typed, guarded container control: ps/logs/inspect/exec/start/stop and compose up/down with JSON output, project-aware targeting, and approval-gated destructive ops.

Thin on-demand context retrieval: context_query / context_slice / context_grep tools that read already-persisted history back with replay-safe citations.

UltraCode-style multi-agent orchestration: a generatable, savable, governable, observable, resumable workflow layer.

AgentTeams multi-agent teams.

Scheduled coding runs in fresh agent sessions with auditable history.

Detect and repair silent Excel formula errors: pattern validation, semantic Formula IR compilation, deterministic and LLM repair, oracle scoring, and chart validation.

Bring your existing Agent Skills (SKILL.md) library to DSH: discover skills across Claude/Codex/Cursor/Gemini paths, inject a progressive-disclosure index, and load bodies on demand.

Indexes workspace symbols with Tree-sitter and provides lexical or optional embedding-assisted code search.

Scheduled agents on a cron: run a prompt on a schedule and get the digest where you already are, with overlap/missed-run/timeout safety defaults.

Adversarial checkup → fix → review loop toolset.

Embedded task management engine: decision-point protocol, idea capture wall, Linear-style issue store.

Pair a second model that passively reviews each turn and injects notes.

Adds specification artifacts, skills, commands, goal-backed implementation, and task-progress context.

Independent read-only acceptance layer: spawns a read-only verifier before each top-level turn closes and steers non-pass gaps back into the agent.

Open DSH workspace directories in VS Code directly from the web GUI.

Desktop notifications for turn completions, with per-outcome controls and keyword rules.

ACP bridge between BitFun and DSH.

ACP profile plugin and standalone stdio server for using the full DSH agent from Zed and other ACP clients while sharing DSH credentials and sessions.

Bridge to the Telegram Bot API: long polling, per-chat sessions, HTML formatting.

Chat with, monitor, and approve your DSH agents from WeChat via the iLink gateway: text both ways, session targeting, digest heartbeats, and numbered approval prompts.

Vision bridge for text-only models: paste an image, get structured JSON evidence (OCR, layout, semantics).

Find plugins without leaving the agent: search this curated registry by keyword or category, with ready-to-run install commands.

Adaptive deep-research orchestrator built on the official workflow engine.

An MC-Fabric-style hook processor.

Pin Git commits to the environment's own author identity; env-var injection overrides all `git config` settings.

Plugin health checks: manifest protocol / patch format / build traps, zero-dependency and read-only.

Local security audit: config, plugin origins, sessions, network exposure — read-only redacted risk report.

Frame-level scan diagnostics for session files (torn/corrupt/empty detection).

Self-evolution: the agent hot-mounts/removes persistent plugins on itself mid-session.

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

Support for the microsandbox backend.

Microsoft cross-platform sandbox support.

Support for the nono sandbox backend.

Agent-tree token budget management.

Role-based LLM retry & fallback strategies.

The model switch for DSH: point it at any OpenAI-compatible endpoint, with curated free/cheap DeepSeek provider presets and automatic fallback when a free tier rate-limits you.

DSH meets Kimi WebBridge.

Community distribution: TUI, desktop, and Web UI as one bundle with layered installation.

Select browser elements directly during Vibe Coding and send structured visual feedback to the DeepSeek Harness Agent.

Reusable Markdown prompt profiles for DeepSeek Harness with per-turn model selection, argument substitution, and state restoration.

Native interactive Git diff review for DeepSeek Harness with structured annotations sent back to the current Agent session.

Adds approval-gated Git status, diff, log, commit, branch, and optional checkpoint tools.

Adds dangerous-operation policy checks, output redaction, and a security-review workflow.

Take control of your tmux panes: list/send-keys/capture, run long jobs in a pane with watch mode, and approval-gated destructive commands.

DSH self-updater in the settings page: one-click check/pull (git pull --ff-only), auto background checks, version diff and changelog preview with a red-dot reminder.

/repro exports a minimal, secret-scrubbed, replayable problem bundle: the session log, failed commands, and git diff.

A TPS metrics plugin.

Compile SVG paths and keyframe specs into self-contained Lottie JSON animation files.

Auto-log failed tool calls across native tools, PTC run_code, and inline invocations: dedup and count root causes into a skill so repeated mistakes fade.

A terminal UI (TUI) for DeepSeek Harness.

Deep links: open a specific session or workspace via `?session=` / `?workspace=`.

Short-video sidebar: native player, series navigation, precise history replay.

A Rust/ratatui terminal client that speaks the DSH SDK JSON-RPC protocol directly and runs standalone or as a profile bundle.

Fixes the bug where your account can't lose money while you code.

Automatically add emojis to AI replies.

Side-panel arcade: 18 offline mini-games to play while the model thinks.

Bidirectional sticker reactions between user and agent.

Desktop pet (QQ-pet style): floats in the corner, draggable, feedable, playable.

Append a thank-you note after every message. Mind your manners.

Pops up a mini-game menu (wordle, match-3, extensible) while the model generates.

Auto chess: human vs AI, or AI vs AI.

Settings extensions for DSH.

Web Components support.

DeepSeek account balance and session cost in the composer dock, with auto-fetched official pricing and peak/off-peak support.

Play Gomoku against the AI, or let two AIs battle it out.

Plan review with anchored annotations and structured feedback back to the agent.

Zero-dependency toolkit: time / encoding / json / calculator / csv / regex / markdown / diff / stat / schema — ten deterministic tools in one install.

Claude Science-style research workbench: ReAct research-loop engine (research_* tools), versioned artifacts with provenance (artifact_* tools), and 10 science skills for genomics/pathogens/bioinformatics.

Notifications for four session states, with browser alerts and prompts.

Two-way WeChat (iLink) bridge: turn-end and approval-request push, in-chat approve/reject and message injection, persistent dedup and convergent long-reply chunking; channel layer extensible to other IMs.

Structured comparison and unified diffs for text/JSON/CSV/Markdown.

HTML↔Markdown conversion, GFM table normalization, and TOC generation.

One memory layer for every AI tool and agent: Context Bundle injection, prompt-time recall, MCP tools, and turn-end DSH thread capture.

Edit user and built-in system-prompt sections with live preview.

Xcode AI integration for DSH: 26 Xcode MCP tools (mcpbridge) + Apple platform skills + Xcode Intelligence-style persona (agent preset or global bundle).

Transparent rankings and recommendations for the DSH plugin ecosystem: daily auto-fetched topic data, an open scoring model, and rank/search/recommend tools with a settings-page leaderboard.

The `dsh pm` plugin manager: multi-source search (awesome list + GitHub + npm), install/remove/update per profile, and a doctor audit of manifests, bundle patches, and version drift.

Parody ads in 2005-Chinese-web style: sidebar banners, in-chat feeds, corner popups, and a close button whose hit area is smaller than it looks. All fictional.

Recurring loops: `/loop` command + loop tool + activity status bar.

Plugin template repo (based on the official turtle-ui repo).

Query 64 Chinese exams (高考/考研/四六级/CPA/法考…) with rule-aware date math (2nd-Saturday, 1st-Sunday) and countdowns.

Recurring-habit scheduling from pure date math: next occurrence, range schedules, and overdue advice.

ADHD behavioral coaching skill: task breakdown, overwhelm management, launch rituals, and failure recovery.

Multi-engine reverse image search aggregator: Google Lens, Baidu, Yandex, TinEye, SauceNAO, IQDB, Ascii2d.

Detect and download media from Bilibili/YouTube/Douyin/Xiaohongshu with quality and format analysis.

Condition-driven wakeup: durable file/command/http/process/webhook watches that wake the agent.

Desktop notification reminders.

Context injection audit: token costs of instruction chains / skill catalogs / tool schemas, duplicate and conflict detection.

Manual approval mode ("Manual Mode" / "Ask Mode").

Connect ChatGPT OAuth and OpenAI Codex models to DeepSeek Harness, with opt-in search and image tools.

Turn-scoped “Allow for this task” approvals: automatically allow matching `danger-full-access` escalations only for the current task, then expire.

Qwen multi-modal plugin support.

Evaluation harness for DSH plugins: YAML cases drive real headless agent runs, assert on tool calls, args, results and token usage, with a baseline gate for CI regression.

Customize the "deep diving" thinking status label to anything you like.

Read-only runtime management panel for the official DSH MCP client: connection status, registered tools, errors, and reconnect counts through the /mcp command and a Settings tab, with sanitized display and enable/disable patch suggestions.

OpenPencil design preview and editing plugin.