Plugins

Unified DeepSeek Harness connector: edit MCP servers (cordis.patch.yml) and user skills (~/.dsh/skills) from one Web UI settings page.

DeepSeek Harness plugin: import models with capabilities from an OpenAI-compatible /v1/models listing (gateway capability extensions) and edit per-model capabilities (thinking, vision) on its settings page

Achievement and gamification plugin: unlock badges for turns, tool calls, sessions, daily streaks and coding behavior (edits/reads/tests), with a badge wall, unlock toasts and a ctx.achievements SDK.

Rules, commands & skills manager for dsh: /rules slash command, settings panel with visual rule editing (add/edit/disable/delete, free-zone support), user-defined custom commands with {input} arguments, skill management, and automatic backup & one-click restore of AGENTS.md.

Skill Manager page in the DSH settings panel: browse system / user / workspace / preset skills, expand a skill to its file tree, view and edit files, import skills from a zip, and export or delete them (system skills read-only).

Automatic type-check diagnostics: after the model edits or creates TS files, runs `tsc --noEmit` in the background (debounced per project) and reports errors via the `code_check` tool — grouped by file with line/column, severity and TS error codes, closing the "edit → check → fix" loop.

Enhanced ACP server for DeepSeek Harness: block & reasoning streaming, usage telemetry, model/effort switching, permission presets, session resume/archive, per-session MCP servers, and Zed file/terminal forwarding.

Review Changes panel above the web composer with a VS Code sidebar bridge: send editor selections, drag files/folders as tags, and batch accept/reject every listed file.

Journals every AI file write/edit into before/after snapshots for an inline VS Code review flow (accept/reject hunks, per-file undo, no git).

Combine text, vision, and image-generation APIs into one Mix model with automatic routing: text-only requests go to the chat model, user images and agent screenshots go to the vision model, follow-ups keep using the same session image, and agents can generate or edit images with session-scoped call history.

Per-edit approval for write/edit/str_replace_editor: shows a red/green line diff before the file is touched and asks approve once or reject, with a master switch in Settings.

Google AI Studio-style floating preview for the Web UI: a draggable/resizable iframe preview with aspect-ratio and resolution presets, plus a VS Code-style code window with direct editing, confirmed writes, and session-log recording.

Workspace file tree for the Web UI: multi-tab preview (text/DOCX/PDF/Markdown/CSV/images), inline Markdown editing, a Git changes panel, and file operations (rename/delete/copy/move/create) via context menu.

Session-scoped Markdown writing pad docked beside the conversation, with edit/preview modes, selection-based AI rewrites, diff review, undo, and draft recovery.

Skill manager for the dsh web GUI: a sidebar entry plus a panel listing the skills visible to the current session (split into global / local groups) with per-skill on/off switches, create / import / edit / delete. Disabling a skill removes it from the mod

Security-audit methodology skill pack plus the plugin_vet supply-chain gate: eight agent skills (secret scan, dependency audit, supply-chain review, prompt-injection review, audit orchestration, threat modeling, vuln intel, incident response) in Chinese and English editions, with an npm provider bundle that mounts the skills and registers the automated plugin_vet pre-install scanner.

Stable, editable PPTX generation for AI agents — semantic IR in, native DrawingML out

Self-editing desktop workbench for DeepSeek Harness: a full dsh web replica with an edit mode in which the agent modifies the workbench (patch layers, plugins, UI) with hot reload.

Render Markdown to a standalone, self-contained HTML page: an `md_html_render` tool that works in a headless profile, plus a web drawer to browse, preview, edit and export local `.md` files; both share one renderer, with no runtime dependencies.

DSH 上的类 VSCode 编码体验：Monaco 中央编辑器（文件页签/QuickOpen/状态栏）+ Agent 编辑差异审查（整文件差异/采纳/拒绝/归档/回滚），状态持久化到工作区旁车

Forge quality gates for DeepSeek Harness (dsh): task gates, read-before-edit, bash hazard interception and quality scoring, driven by the forge CLI through DSH's typed interception points.

DSH workspace folders: categorize workspaces into folders in the left sidebar (drag reorder for folders/workspaces/sessions, inline edit, one-click expand/collapse, search) plus per-workspace token usage and cost stats (inline totals, hover detail card, o

Recall (撤回), edit-and-resend (编辑重发) and regenerate (重新生成) for DeepSeek Harness conversation messages — Web and Desktop

Third-party provider reasoning-effort for DeepSeek Harness: thinking levels declared per model, auto-adapted from a model knowledge base + wire-protocol inference, edited right inside the official Models page card.