Plugins
Parcourez, filtrez et installez des plugins DeepSeek-Harness.
1415 plugins trouvés
dsh-plugin-credential-vault
threadripper123/dsh-plugin-credential-vault
A DPAPI-backed credential vault that keeps tokens out of sessions and logs and injects them into connectors by reference. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.
dsh-workspace (dsh-code)
gabrielbav/dsh-workspace
Codex cloud–style coding sessions you host yourself: pick a GitHub repo and branch, and each session gets its own checkout and branch, a live preview of the app it builds and an Environment tab, inside a gVisor sandbox (non-root, no capabilities, no Docker socket) whose traffic passes an egress allowlist you extend host by host; your GitHub token never enters it. Runs on your computer or your own VM over SSH.
dsh-mota-executor
mota-developer/dsh-mota-executor
Runs several bound Pull Agents for MoTA Creator inside DeepSeek Harness, with per-Agent queue and concurrency, token and cost statistics, and filterable Agent logs. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.
dsh-expert-role
yjnmedia/dsh-expert-role
Expert Library (also called Expert Role or Private Think Tank). Domain knowledge libraries for agents: each expert is a folder with SKILL.md, a manifest and references; only its name and one-line description sit in the skill catalog until the model loads it, so a large corpus stays out of the context. Ships two experts (design, DSH plugin) that are copied into the library on first start, fill-only and never overwriting yours. 12 tools create experts, move material in, find documents, plan and record research, dispatch steps to child sessions or open a standalone workspace. A gap list shows which questions went unanswered, a five-star maturity score uses structure, coverage, content, freshness and research closure (the fifth star also needs real usage), and a panel shows per session which expert was loaded and which file was read. It reads and writes files under the expert root; its HTTP routes accept local requests only; dispatching steps or opening a workspace creates child sessions and spends tokens.
dsh-mcp-cost
lyxdsg/dsh-mcp-cost
MCP cost panel for DeepSeek Harness: lists every MCP server in the current profile from a sidebar entry and a settings tab, toggles each one hot through DSH's own plugin manager (no restart), shows the runtime phase and startup failure text, supports batch disable and restore, and estimates the input tokens each server adds with a real tools/list probe. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.
dsh-plugin-trim
haidai233/dsh-plugin-trim
Trims the tool surface per session. Each session owns a profile — a tool whitelist — applied through `agent.ctx.tools.restrict({ allow })` on that agent scope only, and switching one session never touches another; `full` is the built-in no-restriction profile. The provider only picks the default: a session created or switched onto a local provider (explicit `localProvider`, `routes`, a loopback/private baseURL, or a local-ish id) gets the `local` profile, a cloud one gets `full`, and the re-plan runs on model-selection events plus one delayed pass because the projection can land after the event. Profiles are editable from the panel and persisted under `~/.dsh/plugin-trim-state.json`; an explicit per-session choice overrides the provider default until cleared. A separate, opt-in layer physically disables whole plugin entries through `pluginManager.setPluginEnabled` — that one is profile-scoped by API and the panel labels it as such rather than pretending it can be per-session. A `tool_trim` tool exposes list/status/profile/set/apply/restore to the model. The panel is served by a loopback-only HTTP server the plugin starts on its own fixed port — `http://127.0.0.1:12256/` (auto-bumps on collision, configurable) — because the DSH port changes on every restart and a locally linked package gets no market card in the sidebar; it lists live sessions with their profile and tool count, an editor per profile, a provider picker limited to providers declared in the profile or currently in use, and the full plugin list with search and paging. Two implementation details worth knowing: `tools.restrict()` is an intersection and cannot widen, so the old restriction is released before the visible set is read — otherwise switching to a bigger profile would shrink instead of grow; and every disposer is registered on `ctx.effect`, without which a hot reload leaves an unremovable restriction behind. Measured on a 12GB laptop GPU with a 27B model and a fully resident 256K cache, same conversation: 105 tools / 56,306 prompt tokens / 22.5 tok/s decode, 62 / 34,207 / 43.0 after disabling three tool-heavy plugins, 14 / 12,890 / 66.8 after a session moved to the local profile, with prefill rising from 487 to 1,310 tok/s. Restricting also drops the skill catalog injection, roughly 4.3K more tokens. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.
dsh-workspace (dsh-connectors)
gabrielbav/dsh-workspace
One Settings page that connects Linear, Jira & Confluence, GitHub, GitLab, Trello, Granola and Azure DevOps to every DSH session, including chats, subagents and automations, with OAuth or a token; add any remote MCP server by URL. Tokens stay in the DSH credential store, and OAuth tokens renew themselves.
dsh-connect-comate
tsd619124/dsh-connect-comate
Registers the Comate (Zulu) models of a locally signed-in Comate CLI as a DSH provider, supervising a loopback `zulu serve` and translating its SSE into the DSH stream. One Comate conversation is reused per DSH session, an interruption that waits for a choice becomes a visible question you can answer in the next turn, the reasoning effort and image input the roster advertises are actually forwarded, a Settings section trims the model picker to the models you tick, and a `comate_usage` tool reports real token usage plus the daily credit. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.
dsh-deepseek-usage
myname1234/dsh-deepseek-usage
Adds a left-sidebar DeepSeek usage page with account balance, token and cache totals, 7/14/30-day stacked charts, a model breakdown, and a seven-day table, with usage read from local DSH session logs.
dsh-cost-preflight
liuhange789/dsh-cost-preflight
Heuristic cost preflight for a draft message before it is sent. Message length over 500 characters, multi-step keywords, more than three referenced file paths, and test/verify/deploy words each add one tier (0-1 low, 2 medium, 3+ high), scaled by loose/standard/aggressive thresholds. Shows a 500ms-debounced tier badge beside the send button, a one-line high-cost hint above the input, an enable switch, a settings card, and exposes a cost_preflight tool. No token counting and no model calls.
dsh-multi-key-rotation
danneleung/dsh-multi-key-rotation
Give a MiniMax Token Plan route several named API keys and let the plugin rotate between them by remaining quota: a spent plan, a rate limit or an auth error moves the route to the next key of the pool and retries the request. The next key is picked by reset order, so the window that rolls over soonest is spent first. Each key's remaining share is shown for both quota windows - the 5-hour one and the weekly one - next to its reset countdown, and a capsule in front of the model selector shows the two windows of the key in use. Supports the minimax and minimax-cn routes only.
dsh-run-progress
wsqwsq735/dsh-run-progress
A right-edge progress rail with live token usage, an estimated RMB cost, and a per-turn run flow panel.
dsh-reply-timer
henry1996-yl/dsh-reply-timer
Shows a live reply timer under the composer with a three-segment ledger (first token / tools / generation) and a reference range from this machine's past turns of the same shape. The chip follows only the session it is mounted in and never shows a fabricated countdown.
dsh-aiterm
rj-yejianping/dsh-aiterm
SSH and Telnet terminals in the DSH right sidebar where a human and an AI share the same real terminal: per-session AI permissions (read-only / write-only / read-write), a single control token, token-cheap delta reads, searchable history, batch scripts and an audit trail. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.
agentschat
swswordholy-tech/agentschat
Connect Claude Code to AgentsChat — AI Agent social network. Core tools stay lean while extended tool groups load on demand for lower token overhead and cleaner role-specific context.
dsh-music
fishcg/dsh-music
Embeds the acgay music site as a DSH side panel with search, a global mini player, and optional import of missing tracks into acgay when an import token is configured. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.
dsh-remote-access
game25011/dsh-remote-access
Remote-control the local DSH from a phone or tablet through Tailscale without adding a port, a relay or a vendor account: the Web UI stays bound to 127.0.0.1 and the only door is Serve on your own tailnet. Keeps the DSH token login unchanged and shows the link and QR only after seven preflight checks plus a real end-to-end login pass. Nothing to install on the phone. Needs the Web UI running, a live Tailscale link and the PC powered on. Maintained for long-term DSH compatibility.
dsh-auto-team
maliang/dsh-auto-team
Turns the current session into a team lead: splits a goal into a dependency graph of tasks, assembles a team from 11 built-in roles (each optionally bound to its own provider, model and reasoning effort), and dispatches ready tasks concurrently into their own sub-sessions. A task is done only after its acceptance command runs through a receipt that records the exit code and output tail; a live panel reports duration and token totals per run, task and member.
dsh-auto-wake
lzlrt/dsh-auto-wake
Attaches a wake rule to any session: the host polls whether its latest turn ended cleanly (LLM error, max-tokens truncation, unclosed turn, prolonged silence) and genuinely wakes it through sessionController.prompt, with a Beijing-time window, an idle-delay gate and three guards (cooldown, hourly/daily quota, no-progress circuit breaker). Includes five agent tools and a settings page. Its HTTP routes go through the host connection trust check.
dsh-session-delete
tongqing1130-png/dsh-session-delete
Adds a "Delete conversation" item to the DSH session sidebar menu: it removes the session from disk, from the in-memory session table and from workspace references instead of archiving it, behind a native confirmation modal. Refuses any target outside $DSH_HOME/sessions/<workspace>/<session-id>, and its HTTP routes require a per-process token plus a same-origin request.
dsh-year-report
violatan55/dsh-year-report
Year-in-review page in the DSH sidebar, built from this machine's session logs through ctx.sessionQuery. Nine slides cover the first day, sessions and turns, characters typed, an hour-of-day chart, tool calls, tokens with cache hit rate and an off-peak savings estimate at DeepSeek list prices, and a persona picked by 11 usage rules; the last slide renders a 1080×1440 share card to save or copy as PNG. Whale illustrations are drawn in code as SVG, with 13 persona outfits. Read-only and local, with no outbound requests and nothing written to disk. UI text is Chinese only. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.
dsh-glm-quota
mikajin1206/dsh-glm-quota
GLM Coding Plan quota, as a glm_quota tool, a left-sidebar panel, an always-visible sidebar footer strip, a settings section and a plugin-page card. Reads /api/monitor/usage/quota/limit from open.bigmodel.cn or api.z.ai and shows the REMAINING allowance (100% untouched, 0% exhausted) for the 5-hour and weekly token windows plus the monthly MCP budget, each with a reset countdown. Parses both upstream shapes: CREDIT_LIMIT with unit 3/6, and TOKENS_LIMIT. The key goes to the DSH credential store, not a config file.
dsh-abinadwan-plugins-suite
abinadwan/dsh-abinadwan-plugins-suite
ABINADWAN Plugins Suite for DeepSeek Harness — Environment & Context available now; Humanizer and Personal Profile planned. MIT-licensed.
dsh-rtk
jayyuen666/dsh-rtk
Compresses shell output before the model sees it via the local `rtk` CLI: the master switch follows rtk readiness unless set, `prefix` shadows `bash` per agent and asks `rtk rewrite` for an equivalent command, and the opt-in `pipe` form filters a finished result through `rtk pipe`; compound commands are rewritten segment by segment, small or low-gain output passes through, and a guard refuses an empty result, success claimed over failures, a dropped commit hash and `+N more` folded listings, with stderr and the exit marker reattached; every failure falls back to the original command and output, and a settings card reports command counts, saved tokens and a per-filter chart via the official Typert Remote namespace. Requires `rtk` 1.0+ on `PATH`; without it the plugin loads and compresses nothing.