Plugins

Better streaming text animation for DeepSeek Harness.

Settings page for per-model capabilities: reasoning-effort levels with their wire spellings, input modalities, and token capacities, written as declarations or sparse catalog overrides.

Template for deepseek-harness plugin development.

DeepSeek Harness 独立社区插件商店：发现、安装并提交经过验证的插件、工具与扩展。 | Independent community plugin directory.

File explorer plugin for DeepSeek Harness: file tree, preview, markdown, syntax highlighting, in-panel editing, VS Code integration - DeepSeek Harness 全局文件资源管理器插件

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

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

Skill pool manager in the Settings sidebar: enable/disable, folder batch import with rename-conflict prompts, state-driven one-click DSH-spec check & auto-fix, system/project scope labels.

ShiGuang Memory Engine (SGME) bridge: multi-agent shared long-term memory via HTTP — L0/L1/L1.5/L2 distillation, scenario-based injection, unified search, and proactive care signals (memory_search / wiki_search / signal_pull / signal_claim / signal_ack), installable as `dsh-sgme`.

Local-first four-bucket Token observability: persistent session/provider/model/day dashboards, trends, budget and anomaly signals, public-rate estimates, safe aggregate exports, and explicit opt-in usage or session-trajectory analysis.

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

Projection-driven live DAG for a Session's subagents and durable workflows: status and navigation, deterministic deep-chain layout, fit/pan canvas controls, and current-session node rearrangement.

Plugin coordination layer: plugins declare their capabilities, commands route to the owning plugin without a model call, tool schemas load on demand to cut prompt tokens, run history consolidates into long-term memory while idle, and failures are classified before any retry.

Zhihu search, agent tools, knowledge bases, and usage tracking for DeepSeek Harness.

DSH 原生 Minecraft Agent 插件：每会话独立机器人 + mc_* / mc_kit_* / mc_admin_* 工具 + 单对话看门狗 + 浏览器 UI + 工作区 .whale-craft 记忆

Learning mode that explains each agent step (WIP).

Chrome side-panel browser agent for DeepSeek Harness with tools to inspect pages, navigate, click, type, scroll, and capture screenshots through a local bridge.

Multi-platform research and deep search pipeline: harvest_scout for parallel discovery across GitHub, Web, Twitter, Reddit, Xiaohongshu, Bilibili, and YouTube, harvest_deep_research for automated comprehensive research reports, plus extract, cross-source verification, and credibility auditing.

DeepSeek Harness (DSH) 跨会话记忆插件：对话模型边干边记的 Markdown 项目记忆，支持专题文件渐进式披露、可选后台静默蒸馏、按项目隔离与热更新设置面板。机制对齐 Claude Code auto memory。 | Agent-maintained cross-session Markdown memory: progressive-disclosure topic files, opt-in silent distillation, per-project isolation. Claude Code-aligned.

Encrypted local credentials vault for the Harness: a web settings page and vault_* tools to store, search and copy passwords, API keys, TOTP secrets and card data, with health audits, expiry rotation, imports/exports and read-only/ask access modes.

Multi-user login gateway for the DSH web UI: the first visit creates the admin account, admins add and manage users in the GUI settings panel, ordinary users see only their own conversations, and unauthenticated visitors are redirected to /login.

Per-call cache billing inside the context-meter popover: current step, round and session totals with hit/miss/output tokens and amounts, automatic Beijing working-hour peak/off-peak pricing, per-model rates, and third-party relays billed too.

DSH-native integration layer for the Minta memory engine: composes the official dsh-mcp-client row, registers the minta-memory-governance skill plus a session-start memory prewarm, and ships a Minta agent preset; requires the separately deployed Minta engine (Python/Docker) to provide the 19 MCP tools - not usable without it.

Computer use through an external USB HID device: name a target in plain language, a vision model locates it in a cropped window screenshot, and a Raspberry Pi Pico 2 moves the real mouse and types on the real keyboard. 7 tools. Windows has window listing, per-window cropping and automatic window raising; macOS needs pyobjc and has neither; Linux is unsupported.