dsh-plugin-sentinel: Static Security Auditor for DeepSeek Harness Plugins
In the DeepSeek Harness (DSH) ecosystem, `dsh-plugin-sentinel` is a zero-dependency static audit tool that parses tar archives in memory to detect potential risks in community plugins before installation. It relies on no scripts or build processes, using only Node.js built-in modules to parse `.tgz` files in memory and prevent malicious code from being written to disk. It supports lexical stripping scans and composite rule detection, performing security checks on `package.json`, source code, `cordis.patch.yml`, and tarball structure, covering high-risk indicators such as process execution, outbound network access, credential access, and path traversal. Audit results are classified by severity (pass/review/block), allowing the model to automatically block high-risk plugins before installation.
Read Moredsh-fingerprint-show: Display the Model Fingerprint in the Web UI Metrics Row
`dsh-fingerprint-show` is a DeepSeek Harness client plugin maintained by Bortor, designed to address the difficulty developers face in confirming model deployment versions. It captures the `system_fingerprint` from OpenAI-compatible APIs and appends it as plain text to the end of the metrics line for each AI response in the Web UI. The plugin features zero intrusion, pixel-level style consistency, hover synchronization, and history retention. Developers can install it by specifying the repository directory and verify model deployment consistency in the browser after installation.
Read MoreDeepSeek Harness Plugin: Balance and Usage Monitoring
dsh-balance-and-cost is a client plugin for DeepSeek Harness that monitors account balance, model consumption, and official time-based pricing. Its core features include real-time balance queries, consumption statistics by model and session, pricing that distinguishes peak and off-peak periods, and support for exporting detailed CSV data and hovering to view three tiers of token details. Plugin data is persisted, and you need to restart DSH after installation to use it.
Read MoreDeepSeek Harness Plugin Introduction: dsh-spec-gate (Spec Gate: Ambiguity Scoring and Seed Locking)
The DeepSeek Harness plugin `dsh-spec-gate`, inspired by the Ouroboros philosophy, aims to address the issue of unstable output quality caused by vague requirements. Through three core capabilities—“ambiguity interview assessment,” “spec Seed construction,” and “pre-delivery acceptance gate”—the plugin transforms ambiguous requirements into verifiable specifications and locks immutable constraints, ensuring that drafts meet the required standards. Users need to install and configure the plugin, follow the “clarify–lock–generate–validate” workflow, and achieve requirement convergence and risk control while collaborating with `dsh-cybernetics` to optimize the workflow.
Read Moredsh-skill-lazy: DSH Skill Catalog Lazy Loading and On-Demand Retrieval
The `dsh-skill-lazy` plugin is designed to address the issue of extensive context Token consumption caused by the skills directory in DeepSeek Harness. It uses compact directory injection and on-demand retrieval mechanisms, displaying only brief summaries in each session and retrieving full descriptions via a tool as needed. The plugin also provides MECE checks, cost auditing, and directory sorting based on invocation frequency, helping users optimize skill library management and reduce storage costs.
Read Moredsh-skill-authoring: DeepSeek Harness Skill Development Assistant Plugin
`dsh-skill-authoring` is a plugin for DeepSeek Harness (DSH) that addresses inconsistent skill library documentation structures and the absence of quality standards. It provides `skill_scaffold` to generate standardized skeletons, `skill_audit` to perform audits based on a 9-point quality gate standard, `skill_check_name` to validate naming, and `skill-write-guard` for real-time write protection. The plugin requires no configuration, depends on Python 3.8+, and is open-sourced under the MIT License.
Read Moredsh-memory: DeepSeek Harness Persistent Memory System
dsh-memory is a memory system plugin for DeepSeek Harness (DSH), designed to persist agent context and prevent information loss. It provides core features such as five-way RRF retrieval, confirmation-based writes, Fernet-encrypted storage, session transcripts, and maintenance tools. It supports custom path configuration via Cordis, relies on Node and Python environments, and is suitable for building long-term, structured, and secure agent memory systems.
Read MoreDeepSeek Harness Plugin: dsh-plugin-model-proxy
`dsh-plugin-model-proxy` is a community DSH plugin that addresses the issue of certain models requiring a proxy due to network restrictions while other models remain directly connected. It uses transport-layer routing rules, supports exact, prefix, and wildcard model matching, integrates DSH credentials, supports HTTP/SOCKS protocols, and takes effect in real time. The plugin does not modify the baseURL to avoid masking the upstream, and it provides UI- or file-based configuration. We recommend installing it with profile management to avoid conflicts.
Read Moredsh-runtime-nutrition-label: DeepSeek Harness Runtime Nutrition Label Plugin
`dsh-runtime-nutrition-label` is a community-maintained plugin for DeepSeek Harness designed to address information asymmetry between a plugin’s declared capabilities and its actual runtime behavior. It generates JSON snapshots that include declarations, observations, and inferred evidence by using observation tools and filesystem events. The plugin supports network scanning, metric collection, and automatic integration, and follows a privacy-first model (it does not log credentials or sensitive content). Developers can retrieve snapshots via the `/nutrition-label` command or an API for local diagnostics and compliance verification.
Read Morebitxeno/dsh-github-picker: Referencing GitHub Issues and PRs in DeepSeek Harness
The `bitxeno/dsh-github-picker` plugin is designed to address the inefficiency of frequently switching interfaces when referencing GitHub repository issues or PRs in the DeepSeek Harness (DSH) Web UI. It integrates with the GitHub CLI, allowing users to search, filter, and insert references directly in the input box. It supports the `@owner/repo#number` or URL format and provides draft scanning as well as scroll-based loading. Installation requires running `dsh plugin --profile web add dsh-github-picker` and restarting the service. It relies on the `gh` CLI for local authentication and supports adjusting timeout and cache duration through a configuration file.
Read MoreDeepSeek Harness UI Enhancement Plugin: Model Selector and Interaction Optimization
`bittersmilezzz/dsh-ui-tweaks` is a UI enhancement plugin for DeepSeek Harness. It aims to address usability pain points in the default interface and provide a more flexible interaction experience. Key features include a model inference intensity slider, file upload support via drag-and-drop or paste, convenient request retry configuration, global keyboard shortcuts, and desktop notifications. By extending the interface functionality, the plugin improves the efficiency of model tuning, file handling, and status feedback, making it suitable for users who frequently operate DSH. Before installation, users should review the source code and license.
Read MoreDeepSeek Harness Plugin: Using OpenAI Codex Subscription via OAuth
This document introduces the `dsh-codex-oauth` plugin in the DeepSeek Harness plugin system. It fills a gap in the upstream adapter by providing OAuth credential storage and login commands for Codex authentication. Installation requires addressing pnpm build limitations, and login supports both Web UI and CLI modes. The plugin registers the `codex` provider route, supports models such as `gpt-5.x-codex`, and allows transport and caching policies to be configured via a configuration file. Usage requires a ChatGPT Plus subscription and enforces strict file permissions.
Read Moredsh-plugin-ui-toggle: Managing DSH Plugin Status in the Web UI
`dsh-plugin-ui-toggle` is a client-side plugin maintained by Biobuilder-ai, designed to address the inefficiency of frequently modifying configurations or restarting services to test plugin toggles in DeepSeek Harness (DSH). The plugin provides a visual plugin toggling feature through the Web UI, enabling changes to take effect immediately and be persistently stored. Its core features include managing plugin states directly in the settings menu, exposing a RESTful API, and writing toggle states into a managed code block at the end of the `cordis.patch.yml` file. It does not rely on additional components and supports developers in dynamically adjusting feature modules, although core plugins cannot be disabled.
Read Moredsh-codex-media: Local OpenAI Codex-Powered Image and Document Analysis Tool
`dsh-codex-media` is a local media analysis tool designed for DeepSeek Harness (DSH), addressing the limitation of pure-text models in directly processing images and documents. The plugin provides image description, document analysis, and image generation capabilities, runs using the local OpenAI Codex CLI, and requires no third-party runtime libraries. It must be paired with the `dsh-drop-to-path` plugin before installation, supports multiple transport methods (CLI, Hermes, and API), and enforces strict permission and security validation mechanisms.
Read MoreDeepSeek Harness Personal Workbench Plugin
`bingoogolapple/bga-dsh-workbench` is a personal workbench plugin for DeepSeek Harness (DSH), designed to enhance the user experience. Its core features include personalized banners and chat confetti celebration animations; integrated language study flashcards with gamification mechanics; and a built-in task kanban board supporting weekly matrix and five-column views, manual management, agent-collaborative execution, and scheduled tasks. The plugin is installed via the DSH plugin mechanism, stores data locally, and is licensed under the MIT License.
Read Moredsh-jj-vcs: Jujutsu Version Control Plugin for DeepSeek Harness
The DeepSeek Harness (DSH) plugin `dsh-jj-vcs` addresses version-control bottlenecks in multi-agent collaboration by introducing the Jujutsu version control system. It provides each agent with an isolated workspace and leverages a change-based model to deliver clear attribution auditing, concurrency safety, and operation tracing. Its core features include creating independent changes, explicit permission tiers, controlled publishing, and Windows compatibility. Developers must install it manually and depend on JJ 0.44.0, using the coordinator script to start tasks and verify changes, thereby providing a robust version-management solution at the community level for enterprise-grade multi-agent projects.
Read MoreDSH Knowledge Forge: A Local-First Automatic Wiki and Skill Memory Plugin
DeepSeek Harness (DSH) Knowledge Forge is a native plugin designed to provide local knowledge distillation capabilities for DSH Agent sessions. It converts conversation trajectories into progressively reusable knowledge and separates operational evidence from knowledge. Its core features include local-first capture, session-aware batching, one-time distillation, Wiki-to-Skill evolution, progressive disclosure, and white-box observability. It uses SQLite to maintain runtime state and Markdown files to store Wiki pages and native DSH Skills, enabling structured knowledge to be accumulated from conversations and improving reuse efficiency. The plugin requires DeepSeek Harness 0.1.0-rc.6 and Node.js.
Read Moredsh-conversation-minimap: DeepSeek Harness Conversation List Breadcrumb Navigation
`dsh-conversation-minimap` is a DeepSeek Harness plugin designed to solve the challenge of locating content in long conversations. It provides a Codex-style visual navigation track with marked and highlighted turns; mouse hover or keyboard controls allow previewing conversation summaries and status. The plugin supports adaptive layouts for wide/narrow columns and short/long histories, is compatible with multiple themes and versions, and is highly performant and non-invasive, significantly improving the readability and development efficiency of long conversations.
Read Moredsh-balance-tasks: DeepSeek Harness Balance and Task Count Estimation Plugin
The DeepSeek Harness plugin `dsh-balance-tasks` displays the balance, peak/off-peak billing tier, and an estimated remaining task count based on historical consumption in a real-time floating card in the bottom-right corner. It supports peak/off-peak billing, data persistence, and a RESTful API. Model prices must be configured to calculate a “normalized” task count. Users can install it via a command, and it takes effect after restarting.
Read Moredsh-danmaku: DeepSeek Harness Trajectory Danmaku Plugin
`dsh-danmaku` is a visualization plugin developed for DeepSeek Harness (DSH) that displays trajectory data in real time as danmaku. It supports various events, including tool calls and assistant responses, and provides speed adjustment and edge-docking features. It can be enabled through runtime injection or persistent installation, with settings automatically saved to local storage.
Read MoreDeepSeek Harness Sidebar Session Anchor Navigation Plugin
The DeepSeek Harness (DSH) Web GUI plugin `biggerboy/dsh-conversation-anchors` renders an anchor rail on the left (or right) side of the conversation area to provide a Codex-like navigation experience. Core features include hover preview, multiple style options, keyboard shortcut navigation, and collapsing the thinking process. The plugin supports DSH 0.1.2 and above, improving navigation and reading efficiency in long conversations, and is available on Web platforms only. Installation requires meeting the Node.js version requirements and restarting the service.
Read Moredsh-plugin-manager: Plugin Manager for the DSH Web Interface
DeepSeek Harness (DSH) introduces the `dsh-plugin-manager` plugin for managing plugins via a visual interface. Its functionality is organized into "Custom Plugins" and "Core Components" sections. Custom plugins can be enabled or disabled immediately, and their status can be viewed, while core components only display their mount status. Installation is recommended using the CLI; alternatively, you can manually copy `node_modules` and append the mount configuration to enable hot mounting. Be aware of the risk that disabling the plugin itself and manual installations may be cleared by `pnpm install`.
Read MoreXia Shuo: DeepSeek Harness Creator AI Workbench
"Shuo Xiang" is a plugin for DeepSeek Harness (DSH) designed to turn conversations into a structured creative workspace. It offers home-page project management, support for standardized workflows across multiple content types (such as courses, official documents, and novels), customizable editing workflows, a three-pane workspace with a resource library and knowledge graph, and export/collaboration features. The plugin stores data locally, supports multi-user collaboration and version control, and is suitable for creators seeking systematic writing workflows.
Read Moredsh-conversation-threads: DSH Dialog Message Bar Navigation Track
`dsh-conversation-threads` is a plugin that adds a Codex-style navigation thread to DSH, helping users quickly locate historical messages in long conversations. The thread appears on the right side of the sidebar, includes short lines corresponding to user messages, and supports hover previews, click-to-jump, and scroll synchronization. The plugin can be installed persistently as a Bundle or activated dynamically as a Skill. It is implemented entirely on the client side and poses no privacy risks.
Read MoreKimi WebBridge for DeepSeek Harness
The `dsh-kimi-webbridge` plugin for DeepSeek Harness (DSH) leverages the locally running Kimi WebBridge daemon to extend DSH agent capabilities into real browser environments. It provides 16 tools covering page navigation, clicking, screenshots, script execution, file uploads, and network monitoring, enabling agents to securely control browsers while maintaining login sessions. Installation requires DSH version 0.1.0-rc.7 and Node.js 18+ and can be enabled via a specific command. The plugin supports Windows, and its code paths are compatible with macOS/Linux. In use, an agent can perform tasks such as opening web pages, reading content, and taking screenshots. Additionally, the plugin protects data privacy by processing everything locally and does not rely on any third-party servers.
Read More