Плагины

Менеджер MCP-серверов для DeepSeek Harness: управление MCP-серверами со страницы настроек веб-интерфейса, глобальное хранение в $DSH_HOME/mcp-servers.json, подключение/отключение во время выполнения.

Плагин статуса dsh с частями на стороне хоста и браузера: состояние работы harness (процесс, слушатель, наличие API-ключа, память, аптайм, список плагинов) в виде JSON плюс SSE-оповещения от сервера, отображаемые как значок в заголовке с разворачиваемой панелью в веб-интерфейсе

Компиляция путей SVG и спецификаций ключевых кадров в автономные файлы анимации Lottie JSON.

Типизированное защищённое управление контейнерами: ps/logs/inspect/exec/start/stop и compose up/down с JSON-выводом, таргетингом с учётом проекта и деструктивными операциями, требующими подтверждения.

Formats and validates JSON/YAML/TOML/SQL, with line-and-column error locations.

Alternative WorkflowEngine provider for DSH 0.1.0-rc.7 that runs model-written orchestration in fresh QuickJS/WASM runtimes with a JSON-only host bridge and bounded memory, execution time, and child-agent fan-out.

KiroCrew bridge for DeepSeek Harness: let your dsh agent delegate to a persistent, self-evolving KiroCrew workspace over ACP (JSON-RPC 2.0 over stdio).

Dual-format plugin development guide: teaches agents to build plugins compatible with both the DSH static plugin package (installable via dsh plugin --profile add) and Agent Plugins 1.0 (plugin.json + skills/). Ships as an Agent Plugins 1.0 plugin, a stan

DeepSeek Harness OCR plugin — offline image text recognition powered by PaddleOCR-json (primary) and RapidOCR-json (fallback). 让 DeepSeek Harness 直接识别图片中的文字，无需视觉模型、完全本地离线运行。

Developer quick-reference toolbox for DeepSeek Harness (DSH): 12 themed categories (195 entries) + zero-dependency dev tools (regex tester, JSON format, timestamp, Base64/URL codec, cron generator, line diff). 开发者速查工具箱。

Vendor parameter translation and deterministic JSON repair for DeepSeek Harness: the /translate command maps temperature/top_p/max_tokens/stop/system across 11 vendors, and the tools/post-execute repair layer (plus the fix_json tool) fixes broken JSON too

Local JSONL run traces for DeepSeek Harness sessions

Codex-style artifact preview for DeepSeek Harness (DSH): produced-files row in chat with inline image thumbnails, and a resizable split-screen side preview panel rendering Markdown / source code / CSV / JSON / images / HTML, plus localhost port preview.

Cross-ecosystem aggregation data layer: syncs the GitHub dsh-plugin topic and the npm registry into one open JSON index, queried by model tools, a CLI, and a minimal stdio MCP server.

Agent governance suite: policy-based tool gating (allow/deny/ask with wildcards and priorities), a structured JSONL audit trail, and per-agent token quotas against the host token meter, with state under $DSH_HOME/gov.

Verification toolkit for DSH agents: evidence-based claim checking against workspace files with line citations, config validation (JSON/YAML), and read-only URL/npm/GitHub submission-readiness probes.

SearXNG-backed web_search provider for the web seam: free, self-hosted, key-less metasearch through your own instance's JSON API, with a bundled loopback-only docker-compose example.

Independent browser acceptance testing for agent deliverables: JSON spec in, real-Chromium verdict out.

Render Markdown + JSON into self-contained HTML documents, cards, dashboards, and galleries (CLI + artifact_render tool).

Human-in-the-loop tools (`human_task` / `human_task_ready_check`): pause the agent until the user performs a real-world action or returns an observation (GUI operation, game testing, visual verification, hardware checks), gated by session consent and an AFK presence check, returning a structured JSON result via a web task dialog.

MCP server manager: side-panel with per-status server listing, stdio/streamable-http setup forms, mcpServers JSON import, exponential-backoff reconnect, and tools registered to the model as mcp__<server>__<tool>.

Desktop (native) control for dsh as a stdio JSON-RPC bridge over a Python core: capture the screen, inject pointer/keyboard input, drive accessibility-tree actions, and gate high-risk operations behind required confirmations.

Scans a project's declared build commands — Makefile targets, package.json scripts, just recipes, deno tasks — and exposes them as ready-to-run agent tools.

Persistent model tool-call trace recorder: every tool call is durably written to a JSONL file that survives restarts, queryable via a structured call_trace tool and a callTraceHistory service, with size rotation and an optional floating canvas UI add-on.