Plugins

PR evaluation and improvement consultant: assesses PR success probability based on historical anti-patterns and success patterns, with one-click scoring and actionable suggestions.

Minimal prompt x Web PTC capabilities: one-line persona, run_code SDK, subagents, Ralph, web search/fetch, and Windows Git Bash; the general-purpose workflow tool stays hidden.

Run Stata through DeepSeek Harness: the native Stata AI Skill service auto-starts with DSH (macOS/Windows binaries bundled), so you can run regressions, do-files and econometrics workflows with no manual setup.

Native vision capability extension, using either Zhipu (free) or Qwen-VL (local).

Vision toolkit for text-only DeepSeek: model-invokable `vision` tool, wrapper adapters for deepseek/opencode-go (v4 flash/pro), Antigravity IDE quota (default, flash/pro) / any OpenAI-compatible VLM / Gemini / local Ollama channels, evidence memory with compaction rehydration, content-hash cache, and a bilingual client panel.

Citation tools for agents: exact DOI lookup, Crossref bibliographic search, GB/T 7714 / APA / MLA / Chicago formatting, BibTeX generation, bounded-concurrency DOI validation and health self-check.

Sticker/emoji expression for assistants: emotion-driven auto-matching, library management, AI image tagging, dialects, style mimicry, and chat-based tag refinement.

Manages multiple Jenkins servers and triggers jobs from a Settings page, model tools, or a per-workspace "Run Jenkins Job" entry, with a bilingual host + browser UI.

Evidence-first vulnerability research workbench for DeepSeek Harness: native audit views, 29 OMV tools, Evidence.v1 findings, reproducible runs, Campaign workflows, and isolated PoC evidence.

Bi-temporal memory backed by a SodaMem daemon: recall is contributed to the prompt while each turn is assembled and every closed turn is ingested, with no tool call and no model call on the read path; each recalled fact cites the turn it came from and carries validity dates, so superseded facts stop being retrieved.

Session hotkeys for the DSH Web GUI: switch sessions like browser tabs (positional Alt+1-9, pinned slots, previous/next, new/archive/rename), keyboard model switching via Alt+M, alternate send, Esc focus hand-off, and Alt+W to close or restore a dsh-split-screen workspace — all rebindable with Windows/macOS presets. Verified on DSH 0.1.1 and 0.1.5.

Warm light and dark theme for DSH Web with responsive layout, bilingual settings, safe Markdown preview, per-block copy controls, and no telemetry.

Real-time session token usage and estimated cost in the composer stats row, with pricing fetched from DeepSeek official pricing page.

Yellow one-line notice for a prompt cache miss after the cache was rebuilt; shows re-billed uncached input tokens, cached hit tokens, and TTFT.

Mobile adapter for the DSH web shell: overlay sidebar/details drawers, responsive composer and settings below 768px, zero desktop regression.

Adds a docked-first vertical shorts wall to the DSH Web UI with YouTube Shorts and Bilibili sources, keyword rotation, four-corner resizing, and Alt+S/Alt+M shortcuts.

dsh-skin 增强版：为 DeepSeek Harness 提供皮肤切换与自定义背景壁纸。在原版基础上修复了两处问题——① 大图/特殊格式上传无响应（改用 createImageBitmap 直接解码、增加处理中与错误提示）；② 主题切换时 overrideTokens 触发 theme/change 导致无限递归栈溢出（加皮肤/明暗守卫断开循环）。壁纸存储从 localStorage 压缩方案改为 IndexedDB 原文件直存，全画质显示、零压缩，并自动迁移旧数据。

Expose DeepSeek Harness agent capabilities as an MCP server, so an external MCP client (e.g. Hermes) can drive Harness to execute coding tasks. Hermes = brain, Harness = arms.

Langfuse observability for DeepSeek Harness: exports session logs as OpenTelemetry traces (GenAI semantic conventions) to Langfuse's OTLP endpoint

Claude Code compatibility layer for DeepSeek Harness: load .claude skills/commands/agents and install plugins from Claude Code marketplaces as native dsh skills.

Evidence-first operating kit: six read-only tools (capability catalog, staged workflow plans, packaged skill reader, bounded local memory search, repository release audit, release checklist) and five packaged skills - plans and audits, never remote writes.

Real-compute verification gates: math simplification, logic proofs, regex validation, E-prover FOL, state-machine checks, and code repair, backed by Go-compiled binaries with prebuilt Windows executables.

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

Bounded, privacy-safe local scanner for Codex, Claude Code, WorkBuddy, and CodeBuddy metadata: atomically aggregates keyed evidence and deterministically advises built-in DSH presets and optional capabilities—without content retention, network access, or preset mutation.