Плагины

Ежечасно проверяет в npm новые версии установленных плагинов и спрашивает, какие обновить.

Граф ветвления диалогов в стиле Git в заголовке сессии: цветные дорожки и кривые форков показывают, какая сессия от какой ответвилась, с переходом по клику.

Самоизменяемая desktop-рабочая среда для DeepSeek Harness: полная копия dsh web с режимом редактирования, в котором агент изменяет саму рабочую среду (слои патчей, плагины, интерфейс) с горячей перезагрузкой.

Backup, restore and machine-migration for DSH: packs skills (which live outside ~/.dsh, in ~/.agents/skills), plugin manifests, core config, memories, todos and optionally sessions into a tar.gz with a sha256 sidecar, then imports it on another machine as a three-way merge against a baseline, listing per-file differences instead of overwriting changes made there. Credentials never enter an archive (filename deny-list of 27 names plus 7 private-key suffixes plus a content gate, enforced again on the import side), every import writes a pre-restore snapshot first, and an archive can be uploaded to any WebDAV server and pulled back on the new machine. Also includes scheduled backups, a read-only viewer for individual files, and an in-panel cloud card; no npm runtime dependencies (packing uses the system tar).

Command resolution forensics: which file a command actually runs, which PATH entry won, what it shadows, how many entries are duplicates or dead, and whether the copies behind one name report the same version.

Test work framework for DSH that scaffolds a product, module and version directory tree, registers seven test_framework_* tools and four runtime skills, and adds a prompt-level gate that reports which requirement-document splits, versions, related documents and environment credentials are still unrecorded before testing starts. Requires the playwright MCP server (no fallback).

Publish size check skill: verifies a deliverable's geometric specs against a target platform before publishing - pixel dimensions, aspect ratio, dpi, file size, format, physical print width and colour mode. Geometry only: it does not review content compliance (QR codes, text-to-image ratio, copyright and prohibited content are out of scope). Pure Python standard library, no third-party packages, and inputs are never modified. Python 3.8+. Covers 10 platforms and 20 slots with per-platform confidence labels - WeChat MP, Xiaohongshu, Bilibili, Zhihu, Weibo, Douyin, WeChat Channels, journal submission, print and client delivery - and resolves a platform by id, Chinese name or alias with fuzzy matching. Two trigger modes: asked directly, or as a parasitic self-check that runs by itself once an image, poster, chart or export workflow has written a file and the user mentioned a publishing target, reporting in one line when everything passes and only expanding when something is wrong. For journal submission and print it judges by physical width (pixels / dpi), which catches files whose dpi metadata passes but whose actual printed size is too small; for multi-image posts it checks whether the whole set is consistent and names the odd one out rather than listing every file.

Skill that reviews your coding-agent session transcripts (Claude Code and Codex only; DSH's own transcripts are not supported) over the last two weeks and reports which skills, hooks and rules work, need attention or go unused. Requires Python 3.

Bundles the kb-curator skill, which keeps a local three-layer knowledge base and refuses to index anything the user has not approved. The L2 retrieval layer needs @kenz1117/dsh-engram installed separately, and the engine runs on Windows only.

Shows which skills each turn invoked and installs, updates or removes skills inside the app.

Semantic code search for DeepSeek Harness: registers two native tools, semble_search (find code by describing its behaviour, returning ranked file:line hits) and semble_find_related (neighbours of a known location), which run the Semble CLI via uvx, plus one bundled routing skill. Requires uv on PATH; the first query on a repository builds its index.

Read-only inventory of every agent preset for the Web GUI, with a matrix of which preset holds which tools, MCP servers and skills, plus CLIs that move a plugin between presets.

Reads the installed plugins, MCP servers and skills, groups them by capability, and reports which preset each belongs in; can register one preset per category on request.

Archives the QQ Zone posts you published yourself into a local library, auto-sorted into three columns — thoughts, dreams and reviews (book/film recommendations) — all three of which can be corrected by hand, one column at a time, and undoing a correction never disturbs the other two. Ships a private web viewer that can be locked with a password, repairs long posts that the upstream API truncates to its first ~500 characters, counts words by year and month, and backs up attached images for offline reading. The cookie is stored on the local machine only.

Lightweight crash-recovery for dsh: after a restart, it detects sessions whose last turn was killed mid-flight and re-injects a continuation prompt that first demands the agent verify real-world state before acting, so side-effecting operations are not blindly repeated. It only reacts to the genuine crash marker (turn/end reason 'interrupted', synthesized on cold read by session repair), so it never misfires on user cancels, rate limits or model errors - which is why it needs no error classification, backoff or configuration UI. Unlike general-purpose auto-continue plugins, it also covers the case where a subagent was interrupted mid-run: the parent turn has already ended normally while it waits, so a naive check sees nothing to resume and the whole subagent tree dies silently; this plugin detects the interrupted child and wakes the parent instead. Single host file, zero external dependencies, seven config constants.

Connects the OpenCode Zen anonymous free lane to DeepSeek Harness: refresh the model list from upstream, tick which models the opencode provider exposes, test any of them with a real request that ends in one of seven verdicts, and order both the settings card and the model selector by a hand-written S/A/B/C capability rating.

Bridges the third-party providers saved in a local cc-switch app into DeepSeek Harness: Anthropic-format tool groups become adapter routes, Codex groups are handed to the built-in pi-ai adapter, and a settings page picks which groups to import and in which order.

Collapses the sidebar global-panel rows to a few visible entries and adds a picker for choosing which ones stay, so a long panel list stops squeezing the workspace and session area; rows are hidden only, never unregistered.

Adds an action to the session menu and the session title bar that opens the artifacts folder of that conversation: it reads the session log, works out which directories the conversation wrote files to, and opens them in Explorer.

Adds an installed-plugin picker beside the composer: drag a plugin name into the prompt at the drop point, open a draggable plugin rail in the right sidebar by double-clicking the entry, and choose which plugins (and remarks) appear in Settings.

Adds F5 / Ctrl+R page reload to the DeepSeek Harness desktop web UI, which production builds lack by default (the "Reload Page" menu item and bare-F5 accelerator are development-gated). Binds soft reload (F5, Ctrl+R) and cache-busted hard reload (F5 with a modifier, Ctrl+Shift+R, appending a _dsh_refresh timestamp) through a capture-phase client key listener.

A floating panel that lights up which HowNet semantic primitives the current answer touches, as a brain-shaped node graph, EEG-style channel traces and a composition card.

A social deduction game played against your own model configurations. A table has N seats (six by default, one of them yours) and some of the rest are played by agents bound to different provider, model, temperature, system prompt and disguise settings; you interrogate them and cross-check the answers to work out which seats are not human. Seat identity and each agent model configuration stay hidden in the host state until a seat is judged, then the disguise parameters are revealed. Every round exports a per-seat conversation JSONL plus the judgment record, and the same seed with the same move script replays byte-identical, so runs with different models or disguise strengths can be diffed.

Three-player LAN Dou Dizhu (斗地主) played from a floating window: one machine opens a room on its own port, the other two join with a six-character room code plus an address. Empty seats can be filled by computer players in four switchable strengths (easy, medium, hard, hell), so one or two people can still play a whole hand; the deal starts as soon as the three seats are filled. 54 cards in four suits, bidding of 1/2/3 or pass, the landlord takes the three bottom cards and holds 20, play runs clockwise, and the first-bid seat rotates clockwise between hands. The rule engine is pure and covers 11 legal shapes, with big joker > small joker > 2 > A > K down to 3, suits never compared, straights, pair-straights and planes excluding the 2s and the jokers, and a rocket that beats everything. The table itself is built to be read at a glance: three seat cards side by side in play order (you, then your next and previous seats) with role, hand count and a bar that runs out, one line that says whose hand is on the table and what would beat it, the bottom cards as text, an in-place play history that folds into a single summary line and unfolds into every trick including the passes, plus hint and auto-play buttons that pick the cheapest legal answer out of the local rule enumeration. Nicknames can be changed mid-hand (twelve characters, duplicates get a suffix). Computers run inside the authority and are handed only their own hand, the public record and the seat counts, so no hand can leak; the authority is the only place that decides which card ids leave it. Computers are driven by the room itself, so a one-person table keeps playing with nobody else on the LAN port. Humans have no turn timer, the collapsed floating pill can be dragged around the screen, and 179 unit tests pin the rules, the wire format, the theme tokens, the drag gesture, the history grouping and the computer seats.