dsh-taskfold: A Context Folding Plugin for Keeping Code Agent Sessions Concise
dsh-taskfold is a memory management plugin for DeepSeek Harness, designed to address context window overflow and high costs caused by the accumulation of historical messages in long sessions. The plugin wraps workflows with the `task_begin` and `task_end` commands, and after a task is completed, it automatically collapses the content into a short title summary while persisting the original content. Users can retrieve the pre-collapse content on demand at any time using the `fold_recall` command. It can be used in conjunction with the built-in compression mechanism, supports installation, and requires a process restart to take effect, helping developers keep sessions organized and control inference costs.
Read Morekilo-zen2dsh: Integrating Free Kilo and OpenCode Zen Models into DSH
`kilo-zen2dsh` is a native plugin for DeepSeek Harness (DSH) designed to simplify user access to free models through Kilo Gateway and OpenCode Zen. Integrated via a native adapter, it avoids extra port and process overhead and supports dynamic model discovery, automatic context-limit calibration, and model filtering. It is easy to install and use, but note that it currently only supports text-output models with tool capabilities, and its free requests are not a Keyless authentication bypass.
Read Moredsh-side-chat: Native Parallel Side Chat in DeepSeek Harness
dsh-side-chat is a plugin for DeepSeek Harness 1.2.1 that provides a native, independent side chat alongside the main conversation. It supports true parallel conversations, drag-to-resize proportions, quoting selected text, and tool calls without cluttering the session list. It requires DSH 0.1/0.2 and Node.js 22+ and is suitable for use cases such as code review and debugging.
Read Moredsh-goal-mode: Visualizing Goal Mode in the DeepSeek Harness Web Interface
The `dsh-goal-mode` plugin introduces a visual Goal mode to the DeepSeek Harness web interface, turning previously hidden command operations into intuitive UI components. It adds a full-lifecycle goal toolbar above the Composer, supporting goal creation, editing, pausing, resuming, and completion, with multiline editing, turn limits, and progress display. A toolbar entry and a dedicated settings page also provide flexible access. Plugin data is stored in real time in the session log and supports optimistic concurrency. Installation requires DSH 0.2.0-rc.2 or later, and the plugin is released under the MIT License.
Read MoreDeepSeek Harness Dock Plugin (dsh-dock)
dsh-dock is a dock plugin for DeepSeek Harness that provides a unified single-panel interface for managing various tools. Its core features include usage accounting, model and balance management, task animations and notifications, runtime status monitoring, fun games, an image understanding proxy, and remote access. To install it, first install the host dependencies and then add it via npm. Users can click the Dock button in the left sidebar and configure unit pricing, push notifications, and chat area display in the popup. Plugin data is persistently stored in the host configuration.
Read MoreDeepSeek Harness Plugin: dsh-academic-paper-search
`dsh-academic-paper-search` is an academic literature search plugin designed specifically for DeepSeek Harness. It serves as a thin adapter layer that reuses the `nature-academic-search` service, providing capabilities such as cross-database search, deduplication, and citation graph retrieval. The plugin requires configuration of environment variables (such as `PUBMED_EMAIL`) and pinned dependency versions to ensure persistence and reproducibility, aiming to simplify the integration of multi-source academic data in agent development.
Read Moredsh-office-com: A Native Windows Plugin for Real Microsoft Office Automation (DSH)
`dsh-office-com` is a Windows-native Office automation plugin maintained by wly8691-jpg. It controls real Excel and Word instances via COM, addressing the limitation of conventional plugins that cannot execute VBA macros or recalculate formulas. It provides 29 low-level tools covering application management, read/write operations, macro execution, and core accounting features (such as journal entries and general ledger); it also encapsulates 6 task-level tools and 8 kb_* knowledge base classification tools, supports three execution modes—preview, managed, and attached—and is suitable for scenarios such as accounting report generation, bulk document processing, and data health checks.
Read Moredsh-gsv-tts: DeepSeek Harness Text-to-Speech Plugin
The DeepSeek Harness voice reading plugin `dsh-gsv-tts` supports dual-mode speech synthesis: a cloud-based Edge mode for zero-configuration quick use, and a local GSV mode for offline voice cloning. Core features include one-click reading, automatic reading control, visual sound settings, voice marketplace downloads, and one-click engine start/stop management. It can be installed via the command line and allows one-click switching between the “Quick Experience” and “Local Professional” modes. Please note that Edge mode requires network access, while GSV mode requires a Python 3.10+ environment and may take additional time to load on first use.
Read Moredsh-tap: A DeepSeek Harness Plugin for Integrating CodeBuddy/TRAE/Qoder Subscriptions
`dsh-tap` is an unofficial DeepSeek Harness (DSH) plugin designed to connect subscription quotas from CodeBuddy, TraeWork CN, and Qoder CN to DSH. The plugin bridges credentials via a local loopback gateway and supports multi-channel model access, tool integration, and usage metering, with credentials stored locally only. Installation requires running specific commands and logging in with credentials. This tool is suitable for users who want to leverage existing subscription quotas to develop agents, but server-side failures and tier limitations should be noted.
Read Morestaff-os/dsh-workbench: DeepSeek Harness Enterprise Workbench Plugin
`staff-os/dsh-workbench` is an MIT-licensed, enterprise-grade workbench plugin for DeepSeek Harness (DSH). It integrates AI employees, knowledge bases, skills, MCP servers, and plugins into a unified local management interface via mount points. The plugin supports local skill directory management, skill marketplace installation with security scanning, and local BM25 knowledge-base retrieval without requiring an external backend. Configuration relies on environment variables and patch files; skill changes take effect in real time, while MCP and plugin changes require a restart. It is suitable for users who need unified management of multi-domain resources in DSH and localized closed-loop operations.
Read Moredsh-auto-compact: Enhanced Auto-Compact Defaults for DeepSeek Harness Agent Presets
`dsh-auto-compact` is a “workflow” category plugin for DeepSeek Harness (DSH), designed to optimize the default context-window auto-compaction configuration when agent presets are running. Maintained by developer `songoao25`, it improves context resource management efficiency by adjusting compaction strategies. The official documentation does not provide a ready-made installation command, so installation guidance must be obtained from the repository. When using the plugin, run it with the current process permissions and review the source code and license to ensure compliance.
Read Moredsh-mihome: Mi Home Smart Home Control Plugin
dsh-mihome is an open-source plugin that provides a Mi Home device control interface for DeepSeek Harness. It allows agents to securely query devices, read status, and execute commands, with all state changes requiring manual approval. The plugin supports connection checking, device listing, status reading, control operations, and dashboard rendering, and is available in demo mode and cloud connection mode. After installation, it must be enabled through DSH configuration, ensuring account security and adhering to the principle of least privilege to provide a secure infrastructure for agents to operate physical devices.
Read MoreWhaleNovel: A Native AI Novel Creation Workbench Plugin for DeepSeek Harness
Write Whale is a native novel-writing plugin for DeepSeek Harness, designed to address context fragmentation and scattered assets in long-form creation. The plugin integrates outlines, main text, and Agent workflows, providing a WYSIWYG workbench, book-library asset management, version control, Agent collaboration, and review optimization features. It supports drag-and-drop chapter reordering, autosave, and a ChangeSet proposal mechanism, ensuring users retain full control over all AI edits. The plugin includes built-in strict review and NOAI scanning capabilities, along with a self-evolving mechanism that adapts to the author's style. Novel content is stored locally, with data security assured. It is suitable for authors who need AI-assisted continuation, rewriting, and review, and who want to retain Markdown-based management capabilities.
Read MoreDeepSeek Harness Scheduled Task Plugin (DSH Automations)
dsh-plugin-automations is a scheduled-task plugin for DeepSeek Harness Web Profile. It provides one-time execution, Cron-based scheduling, and two execution modes: "on time" and "off-peak hours." The plugin supports custom Session titles, status persistence, and a 30-minute timeout. Tasks are submitted via the HTTP API. It is currently in the MVP stage and supports task creation, but does not yet support editing or deletion. It requires a Node.js 20+ environment.
Read Moredsh-session-recycle-bin: DeepSeek Harness Session Recycle Bin
`dsh-session-recycle-bin` is a DeepSeek Harness plugin designed to address session accumulation. It provides a standard recycle bin mechanism, supporting quick sidebar deletion, bulk management via the settings panel, and true physical permanent deletion. With features such as active session protection, batch fault tolerance, and adaptive UI, it standardizes session lifecycle management and helps prevent accidental deletion of important data.
Read Moredsh-session-persistence-mysql: MySQL Persistence Backend for DSH
The `dsh-session-persistence-mysql` plugin provides a MySQL-based session persistence solution for DeepSeek Harness. It implements the `SessionPersistence` interface and supports transactional appending, tear-free tail handling, crash recovery, and read/write separation. The plugin is configured through environment variables and supports lazy instantiation and a deadlock retry mechanism. Known limitations include no support for deletion or paginated filtering, and the use of plaintext storage. It is suitable for production environments with high data reliability requirements.
Read Moredsh-ctrl-enter-submit: Change DSH submit shortcut to Ctrl/Cmd+Enter
`dsh-ctrl-enter-submit` is a DeepSeek Harness (DSH) plugin designed to resolve the conflict of default Enter-key submission during multiline input. By intercepting keyboard events, it changes the submit shortcut to Ctrl/Cmd+Enter while preserving Enter for line breaks. The plugin supports the main dialog and Agent question cards, and can be enabled or disabled in the settings after installation. It is compatible with Windows, macOS, and Linux, requires dsh web version 0.1.0-rc.6 or higher, and does not interfere with input methods or menu operations.
Read Moredsh-mnemos: Controlled Cross-Session Memory Plugin for DeepSeek Harness
The DSH plugin dsh-mnemos is purpose-built to address agents' "context loss" and disorganized memory management. It is a controlled, extensible cross-session memory plugin that ensures data security through a human approval queue and provides memory retrieval, batch distillation, skill transformation, and SQLite/Git-based storage synchronization. Supporting an open bus mechanism for third-party integration, it also offers a `/memory` command and API interface, aiming to build a long-term memory infrastructure for DSH agents.
Read MoreDeepSeek Harness Plugin: OpenCode Go Model Sync
dsh-opencode-models is a plugin for DeepSeek Harness that addresses stale server-side model lists. It uses live probing—text and image requests—to fetch the latest OpenCode Go models, verify availability and vision capability, and intelligently update the configuration. The plugin includes multiple safeguards, such as preserving manually added metadata, avoiding hard-coded context windows, and aborting updates when a large number of probes fail. It is easy to install, compatible with dsh ≥ 0.1.0-rc.7, and supports custom Provider configuration and thinking intensity.
Read Moredsh-graphify: Native Code Knowledge Graph Plugin for DeepSeek Harness
The `dsh-graphify` plugin for DeepSeek Harness (DSH) natively connects to a local knowledge graph engine, providing agents with session-aware project parsing, code change tracking, and dual-mode (Compact/Full) graph query tools. It supports incremental updates and full builds, features local runtime resilience, and helps agents construct structured code maps. Installation requires Node.js, Python, and uv, and has been tested on macOS/Ubuntu; Windows is unverified. The `/graphify` command builds or refreshes the graph, `update` processes only code changes, and documentation changes require a full rebuild.
Read Moredsh-noname-kit: Noname Kill Extension Development Workshop
`dsh-noname-kit` is the “Noname Extension Development Workshop” plugin for DeepSeek Harness. It encapsulates the development process into spec injection, official source-code search, static validation, and a dedicated write channel, addressing the pain point of developers having to manually handle many low-level details. The plugin supports Web and desktop installation. Its “Workshop” tab provides task management (create, edit, and archive) and P2P online testing. It also includes a built-in cross-network online assistant that supports automatic extension transfer resumption and kernel management. Developers should be mindful of environment differences and use the dedicated `noname_write_extension` tool instead of regular write commands to avoid rejection by the plugin guard.
Read MoreDeepSeek Harness Novel Studio: Turning AI Dialogue into a Creative Workbench for Writing Full-Length Novels
Novel Studio is a long-form novel writing plugin built specifically for DeepSeek Harness (DSH), aiming to address issues such as fragmented settings, broken logic, and disorganized character relationships in creative writing. It separates conversation from project management and provides core capabilities including structured outlines, character relationship trees, worldbuilding maintenance, and consistency validation. Supporting imports in multiple formats and adopting local-first storage, optimistic revision, and human confirmation mechanisms, it ensures data security and the rigor of creative logic, making it well suited for users who prioritize consistency in long-form writing.
Read Moredsh-honcho: Providing Persistent Memory for DeepSeek Harness
dsh-honcho is a DeepSeek Harness plugin maintained by Plastic Labs that provides persistent memory through Honcho, addressing context loss issues. It supports injecting context at session startup, background message capture, and built-in tools, and offers status viewing. Installation requires a Node.js environment, adding the plugin via commands, and setting the API key, companion name, and a shared workspace value in the configuration file to enable memory sharing.
Read Moredsh-workspace-mover: True Cross-Workspace Migration of Sidebar Sessions
The DeepSeek Harness plugin developed by PianoPrince complements native cross-workspace session migration. It supports moving sessions via drag and drop or tool calls, and offers features such as batch migration, group merging, zero-token true migration, a workspace moving wizard, session rescue, and recycle bin backup. Written in pure JavaScript source code, the plugin provides move-history undo and task-center management, and must be installed under Web Profile with a restart. After installation, it can be operated in the grouped view, supports multi-selection with Ctrl/Shift, but cannot move running sessions.
Read Moredsh-context-mode: DeepSeek Harness Context Window Management Plugin
`dsh-context-mode` is a DeepSeek Harness (DSH) plugin designed to solve context overflow issues caused by large files and web content through the “everything is a plugin” philosophy. It primarily features three capabilities: local knowledge base indexing (based on SQLite FTS5, with support for web scraping and chunked retrieval), routing interception (hard-blocking large-file tools and soft-guiding precise reading), and sandboxed execution (running code in an isolated environment and returning only the conclusions). By optimizing data inputs, indexing, and isolated execution, it effectively reduces context window pressure and improves the efficiency of context utilization.
Read More