dsh-notifications: Desktop Notification and Alert Sound Plugin for DeepSeek Harness Web
`dsh-notifications` is a client-side plugin for DeepSeek Harness Web, designed to address the problem that background-running tasks are easily overlooked. It uses native browser capabilities to provide desktop notifications and alert sounds, monitoring thread status changes (such as completion, cancellation, and approval) to trigger reminders. The plugin is purely front-end, has no external dependencies, supports red-dot indicators on tabs, automatically focuses the window when a notification is clicked, and includes multi-tab notification delivery priority logic. After installation and enabling, users must grant notification permission to unlock alert sounds. It is suitable for task scenarios that require long-running background execution or timely response.
Read Moredsh-upload-plugin: DeepSeek Harness File and Image Attachment Plugin
The DeepSeek Harness plugin `dsh-upload-plugin` addresses the lack of attachment functionality in the native editor. It allows users to upload images (as native drafts) and files (via context injection) directly through toolbar buttons or commands, without manually pasting paths. Additionally, the plugin supports browsing upload history, configuring the model's Vision capability, and setting the reasoning level. Installation requires running the command under the `web` profile, making it suitable for scenarios where frequent interaction with documents or images is needed during conversations.
Read Moredsh-plugin-brief: Convert Agent Responses into Briefing Slides
dsh-plugin-brief is a DeepSeek Harness plugin designed to convert the agent’s technical output into an easy-to-read slide view. It uses local typesetting by default, offers a “plain-language” rewrite feature, and incurs costs only upon user click. The plugin hides distracting information while retaining key steps and pending issues, supports switching between the original and plain-language versions, and allows viewing the full original text. It is simple to install and can be enabled via the `dsh plugin` command.
Read Moredsh-queue-reorder: Reorder Messages in the DSH Queue Dock Bar
`dsh-queue-reorder` is a client-side plugin designed for the official DeepSeek Harness (DSH) queue dock to address the inconvenience of adjusting message order. It adds grip drag handles and ↑/↓ fine-tuning buttons, allowing users to reorder queued `next-turn`-type messages directly in the interface. The plugin uses a non-invasive decorator design, preserves existing functionality, and has zero dependencies, so it works immediately after installation. However, version compatibility should be considered, and because it depends on internal DOM structures, official refactoring may cause the interactions to break.
Read Moredsh-session-notice: DSH Session Bark Notification Plugin
dsh-session-notice is a web-connected tool plugin for DeepSeek Harness, used to monitor agent session states. Users can enable a separate notification status in the session interface via the bell button; each time a turn stops, the plugin pushes a message to the user's phone through the Bark service. The article details installation and configuration (including server and key setup), usage steps, and behavioral rules (including message summarization logic), and provides technical details and precautions.
Read MoreDeepSeek Harness Plugin: dsh-wechat
The `dsh-wechat` plugin for DeepSeek Harness (DSH) bridges WeChat contacts into DSH, enabling AI services to be invoked within WeChat. Core features include independent sessions for each contact, support for text, image, and file transfer, and native typing indicators. Access for specific WeChat IDs is controlled through a whitelist mechanism, and the login process is executed locally. During installation, configure `allowFrom` and restart the service. Notes include concurrency limits, lack of voice and video support, and the need to log in again when a session expires. It is suitable for scenarios where WeChat is used as the entry point for AI interaction.
Read Moredsh-retrace: DeepSeek Harness Session Time Machine
dsh-retrace is a DeepSeek Harness plugin that adds “time machine”-style conversation tracing to the web interface. Built on a versioning mechanism, it supports one-click rollback of an entire conversation turn, editing and resending historical prompts, response regeneration, and rollback of conversation context and file artifacts. Developers can install it via the command line and use it after a restart. The plugin mitigates issues caused by mistaken operations and difficult iteration, enhancing the security and controllability of DSH conversation management.
Read Moredsh-font-customizer: DeepSeek Harness UI and Code Font Customization Plugin
DeepSeek Harness (DSH) introduces the dsh-font-customizer plugin, designed to address users' needs for personalized UI and code fonts. The plugin allows users to independently configure UI font stacks and monospaced code fonts, and provides fine-grained body font size adjustment from 12 to 17 px. Its core strengths include local font detection and scanning, built-in presets for popular fonts, and customization through CSS font stacks. Settings are written atomically to local storage, enabling automatic cross-device synchronization and one-click reset.
Read MoreDataTally: DeepSeek Harness Data Asset Usage Signal Aggregation Plugin
DataTally is a plugin for DeepSeek Harness that aggregates public usage signals of data assets, such as download counts and Star counts, to provide AI agents with tools for querying and comparing data asset usage. The plugin focuses on recording and presenting facts, without performing value calculations, rankings, or recommendations. It includes search, resource retrieval, and comparison features, supports both CLI and Web contexts, and helps developers objectively assess the activity level of data assets.
Read MoreDeepSeek Harness Version-Locked Update Plugin dsh-lockstep
`dsh-lockstep` is a plugin in the DeepSeek Harness (DSH) ecosystem for strictly enforcing consistency in dependency versions. It addresses the risks of DSH’s default “floating” update strategy during release-candidate stages or in high-privilege environments by comparing local lock files with remote state and planning pinned upgrades rather than automatically updating to the latest versions. Users need to run `check`, `plan`, and `apply --write` in sequence to apply changes. The tool can be installed standalone or used as a DSH plugin, includes safety mechanisms, updates configuration files only without executing installation scripts, and supports bypassing API rate limits via `GITHUB_TOKEN`.
Read Moredsh-turn-doctor: DeepSeek Harness Turn Failure Diagnostic Plugin
DSH-turn-doctor is a diagnostic plugin for DeepSeek Harness that helps identify the causes of failures such as request timeouts. It measures time to first byte, byte inter-arrival gaps, and total request time, analyzes multi-layered timeouts across idle watchdogs, SDKs, HTTP timers, and tool calls, pinpoints the specific failure point, and provides remediation suggestions. The plugin is strictly for diagnostics: it does not perform retries or modify configurations, and its verdicts remain in memory, helping developers avoid blind configuration changes and resolve issues quickly.
Read Moredsh-model-pin: Limit model request scope to prevent llama.cpp router reloads
The DeepSeek Harness (DSH) plugin dsh-model-pin resolves the “model roulette” problem, preventing local model reloads, cache loss, and routing errors caused by requests to disallowed models. The plugin enforces an allowlist at the resolution layer, forcing mismatched requests to be redirected or rejected, ensuring that sub-agents and background tasks always use the specified model. It is suitable for resource-constrained environments and scenarios requiring strict routing control, and can be easily enabled and customized via configuration files.
Read Moredsh-logbook: A DSH Plugin That Makes ctx.logger Output Visible
`dsh-logbook` is a DeepSeek Harness plugin designed to address the issue of invisible default logs. It provides a `/logs` command and a stderr exporter to make the output of `ctx.logger` visible. It supports filtering logs by level and name, and allows flexible configuration of log retention counts and muting specific loggers in `cordis.patch.yml`. This plugin resolves the bug where buffered warn/debug logs are silently discarded, making it a practical tool for troubleshooting in the terminal.
Read Moredsh-context-budget: A Plugin for Keeping Local Model Context Within a Controllable GPU Range
The `dsh-context-budget` plugin is designed to address the sharp performance degradation of local models under long contexts. It measures prefill speed and enforces hard limits and checks for token count, first-token latency, and cold prefill time. When a limit is exceeded, it can print a warning or perform compaction, keeping the context within the range the GPU can process. The plugin coexists with `compaction-basic` and provides a `/context-budget` command to monitor status, helping users optimize the experience of using local large language models.
Read Moredsh-protect-eyes-skin
The eye-protection theme plugin `dsh-protect-eyes-skin` for DeepSeek Harness (DSH) helps reduce developers' visual fatigue with a rose green, low-blue-light color scheme. It is a pure front-end implementation with zero runtime dependencies, covers all 79 design tokens, automatically follows system settings for light and dark modes, and supports clean uninstallation. Users can install it via the command line and activate it by switching the appearance in DSH settings.
Read Moredsh-llm-xai-oauth: Using a SuperGrok / X Premium Subscription in DeepSeek Harness
The DeepSeek Harness plugin integrates SuperGrok/X Premium subscriptions into DSH by reusing local tokens, supporting the Grok 4.x model series. Installation requires Node.js ≥ 22.19, can be completed via `dsh plugin add`, and includes OAuth device-code login. The plugin supports automatic token refresh, but a daemon or scheduled task is required to keep it running. To use it, simply specify `provider xai` and `model grok-4.7` in the configuration to invoke it.
Read Moredsh-self-memory: DSH Self-Memory Plugin
The `dsh-self-memory` plugin is designed to address the lack of cross-session experience reuse for DeepSeek Harness agents. By recording session trajectories and solutions, it provides offline retrieval and closed-loop write-back capabilities, improving the efficiency with which agents solve recurring problems. The plugin offers five core capabilities: multi-keyword association, generalized summarization, fast retrieval, premise filtering, and synonym normalization. It supports hot-mount installation, with automatic collection and injection of historical solutions enabled by default. Users should follow a closed-loop workflow: “check records first → try a match if found → record the root cause on success or failure → develop a new solution if none exists.” The tools include status queries, solution retrieval, record appending, and success/failure marking. Configuration is flexible, allowing the recall budget and storage directory to be adjusted, and it has no external dependencies. Retrieval currently uses keywords and an inverted index; migration to SQLite is recommended when the record volume becomes large.
Read MoreIntroduction to the DeepSeek Harness Process Panel Plugin
`dsh-process-board` is a service management plugin for DeepSeek Harness (DSH), designed to address the challenges of backend service visibility and operations. The plugin automatically identifies and groups the sessions associated with services using environment variables, features a built-in smart filtering logic to display actual service processes, and supports tri-state probing, start/stop management, real-time log viewing, and full-screen mode. Developers can explicitly register service information via Shell environment variables or the API for precise control. The plugin currently supports Windows and Linux, but its process scanning feature is not yet functional on macOS.
Read MoreDeepSeek Harness Plugin: dsh-model-modality
The DSH plugin `dsh-model-modality` addresses the issue of third-party models being unable to automatically detect image input in DeepSeek Harness. It allows users to explicitly declare model metadata (`inputModalities`) to ensure that images, attachments, and other features work correctly. The plugin supports writing configurations for both official and third-party providers, provides declaration and query tools, and offers an optional browser-side UI. The installation command is `dsh plugin add ct-jyjntc/dsh-model-modality`, and it requires a specific version of `@deepseek-ai/dsh-tools`.
Read Moredsh-garmin-coach: A Sports AI Coach Plugin for Garmin Connect
`dsh-garmin-coach` is a local plugin that integrates Garmin Connect data into DeepSeek Harness (DSH). It syncs activity data via reverse-engineered API endpoints, supports authentication for CN and international accounts, and uses AI to generate training recommendations and plans, with all data stored locally in the `~/data/` directory. After installation, bundles must be configured manually and DSH must be restarted. Users should be mindful of sandbox file access permissions and account security risks during use.
Read Moredsh-skill-mcp-bridge: Bridge Plugin for DSH Project Skills and MCP Servers
dsh-skill-mcp-bridge is a bridge plugin designed for the DSH ecosystem, intended to address the cumbersome manual configuration of skill files and MCP server connections. It supports automatic import of skill files such as `.claude`, `.agents`, and `.trae`, and can bridge MCP servers over HTTP, SSE, or stdio protocols. It also provides a GUI toggle panel and state persistence capabilities to improve agent development efficiency. The plugin is maintained by developer cstarc, and users should review the source code themselves before installing it.
Read Moredsh-voice-mini: A Voice Feedback Plugin for DeepSeek Harness
`dsh-voice-mini` is a voice feedback plugin for DeepSeek Harness that uses TTS technology to convert an agent's reasoning process into speech, helping users handle multiple tasks. The plugin supports an autonomous "speak" mode driven by model decisions, status template announcements, random voice assignment per session, and a native macOS floating pet. It provides multiple TTS backends, including cloud-based Edge and local Kokoro, and includes a monitoring panel. Installation requires building the plugin and linking it to the DSH configuration. It supports three presets—instant, default, and minimal—as well as internationalization, making it suitable for developers who want to monitor progress through audio.
Read Moredsh-reasoning-effort-slider: DeepSeek Harness Reasoning Effort Slider Plugin
`dsh-reasoning-effort-slider` is a plugin for DeepSeek Harness, designed to address the issue where reasoning effort is silently turned off when switching models in the native logic. The plugin injects a slider into the chat bar, supports multiple models including Bailian and DeepSeek, and allows independent adjustment of reasoning effort per conversation. Its core features include smart UI display (showing only supported levels), automatic memory of each model’s configuration, and automatic inheritance or graceful downgrading of the remembered effort level when switching models. This mechanism prevents thinking capability from being silently disabled and supports installation with a single command.
Read Moredsh-bailian-models: Alibaba Bailian Model Presets and Automatic Adapter for DeepSeek Harness
`dsh-bailian-models` is a Bailian model plugin for DeepSeek Harness, designed to resolve compatibility issues with the `dashscope.aliyuncs.com` endpoint for native adapters. It provides a preset catalog of 36 models with metadata such as context windows and inference parameters, and includes automatic adaptation that can complete missing dialects, inference modes, and budget settings. The plugin supports four inference modes—Effort, Budget, Toggle, and Always-thinking—and can be used directly by configuring the key via environment variables.
Read Moredsh-mmx: MiniMax CLI Bridge Plugin for DeepSeek Harness
DSH-mmx is a plugin that bridges MiniMax CLI capabilities into DSH. It provides two core features: first, registering the mmx-search provider to enable a web search fallback when a DeepSeek search key is unavailable; second, transparent image recognition, automatically converting images sent to text-only models into text descriptions, while allowing multimodal models to be configured to route image processing through mmx or not. The plugin supports one-click installation and login for the mmx CLI and includes robust security mechanisms (no key logging, temporary file permissions set to 0600). After installation, you need to pin the mmx-search provider in the configuration and enable the feature on the Settings page.
Read More