jaxxchen003/foliodrop-dsh-publish: Publish agent-generated content as FolioDrop share links
## Introduction A common challenge in agent development is the finalization phase: when an agent has drafted an HTML page, a Markdown document, or a code snippet for demonstration, but there is no ready-made channel to share it with others. Whether by manual copy-pasting, setting up your own service, or uploading to a generic cloud drive, these methods all involve extra steps, and once the content is shared, it is immediately detached from the confirmation process. DSH (DeepSeek Harness) philosophy is "everything is a plugin," so such external publishing capabilities are ideally suited to be mounted as plugins rather than being hardcoded into the core. The foliodrop-dsh-publish introduced below does exactly this: it allows you to connect your FolioDrop account within DeepSeek Harness to publish content you have reviewed into a shareable URL. ## What is this
Read Moredsh-tool-visual-primitives: Supplement visual capabilities for DSH text-only models
# Introduction When building agents with DeepSeek Harness (DSH), you often encounter the same limitation: the text model at hand lacks native image input capabilities. If you paste screenshots or data charts into the conversation, the model cannot read them. Without plugins, you can only use a vision model upstream to read the images and manually paste the descriptions back into the conversation, breaking the chain in the middle. DSH's philosophy is "everything is a plugin," and such gaps are suitable for being filled by plugins. The dsh-tool-visual-primitives introduced below is exactly such a plugin: it hands images over to an external vision model for analysis, returns pure-text visual evidence with spatial localization information, and then passes it back to the original conversational model to continue reasoning. The enhanced model does not require native image input capabilities. # What is dsh-tool-visual-pr
Read Moredsh-diary: A Markdown diary for the DeepSeek Harness Agent
# Introduction When building agent applications, a specific issue often arises: while models can reason and invoke tools, there is nowhere to store "what happened today". Once a conversation ends, records are scattered within the chat, making it difficult and passive to review or retrieve them. The philosophy of DSH (DeepSeek Harness) is "everything is a plugin"; such capabilities are best implemented as plugins to fill this gap. The dsh-diary introduced below is exactly such a plugin: it provides a set of diary tools for the model, paired with a browser-based interface for viewing and managing the entries. # What is this dsh-diary is a DeepSeek Harness plugin maintained by hyperMoss. Its npm package name is `dsh-diary`, the current version is 0.1.1, and it is licensed under MIT. It does three things: 1. In `ctx.`
Read Moredeepseek-balance-banner: Top dual-currency balance banner for the DSH Web interface
# Introduction When developing with the DeepSeek API, three questions often require a pause for confirmation: how much balance remains in the account, whether the current time is the off-peak period, and how much was deducted from the last call. Since checking the balance requires a separate API call and peak-valley pricing causes costs to fluctuate based on the time period, relying solely on switching pages and mental calculation is unrealistic. The DeepSeek Web interface features a whale widget in the bottom right corner to view the balance (data sourced from the official API), but if you want to keep your dual-currency balance, peak-valley status, and every deduction in view while coding, deepseek-balance-banner creates a banner at the top of the page. Below, we introduce its features, installation, and daily usage. # What is this deepseek-balance-banner is DeepSe...
Read Moredsh-image-skin: Generate DSH theme color scheme from an image, with a desktop pet
dsh-image-skin is a DSH web-based image skin plugin (MIT, v0.1.0) based on the core philosophy of "everything is a plugin". After users upload PNG/JPG/WebP images, the plugin extracts primary, secondary, and accent colors via k-means clustering to automatically generate a pair of light and dark theme tokens, ensuring the background, panels, sidebars, borders, and text colors follow the image's tone. Meanwhile, the image serves as the page background, paired with semi-transparent panels featuring adjustable opacity and blur (canvas pre-baked). The gallery is displayed as a thumbnail grid, with status saved in `localStorage` for automatic restoration upon restart. The plugin also includes optional desktop pets with five characters (cat, rabbit, whale, dragon, and Shiba Inu), which can greet, hide, and be dragged and dropped. Furthermore, they poll the host API to detect the agent's working status, entering a "thinking mode" and providing a notification upon completion. Installation
Read Moredsh-codex-file-panel: Add a Codex-style workspace file sidebar to DeepSeek Harness Web UI
# Introduction When developing agents with DeepSeek Harness (DSH), the Web UI is the standard entry point. However, tasks such as viewing and modifying workspace files, reviewing Git changes, or launching a temporary terminal often require toggling between the browser, editor, and terminal, resulting in fragmented context. dsh-codex-file-panel addresses this issue by consolidating the file tree, search, editing, preview, and Git review into the right-side panel of the Web UI, while also adding capabilities closely tied to file workflows, such as a bottom terminal. The following sections introduce the features, installation, configuration, and security boundaries in order. # What is it? dsh-codex-file-panel is a community plugin maintained by JesonChou for DSH.
Read Moredsh-krill-theme: DSH Web Deep Sea/Shallow Sea Dual Mode Theme Plugin
dsh-krill-theme is a pure client-side CSS theme plugin for DSH Web maintained by i1j (MIT, v1.0.0). It adopts the philosophy of "everything is a plugin" to replace the skin-changing method of modifying `dist` files or overriding class names. A single stylesheet covers both a dark mode (Deep Sea World, where importance is encoded by brightness) and a light mode (Shallow Sea World, where clarity is encoded), automatically switching according to the system preference. It redefines approximately 95+ tokens solely on the `body` element, without modifying `dist` or relying on component class names, ensuring extremely low upgrade costs. Key features include: a canvas-drawn plankton animation background (driven by ocean currents and supporting `prefers-reduced-motion` fallback), deep-dive work state enhancement during generation, and a jellyfish-shaped parameter control panel. It supports configuration before loading, real-time parameter adjustment after loading, and...
Read Moredsh-voice-call — Let agents initiate calls proactively; it's up to you whether to answer.
dsh-voice-call is a DSH voice plugin maintained by PandaPolo (MIT, forked from Jesse-njx/dsh-voice). The core philosophy is to return control of voice output to the user: the agent possesses an "active voice" that can initiate calls via the `offer_call` tool, allowing users to choose to answer, reject, or call back later; if not answered, playback is strictly prohibited. Incoming calls ring via a floating card overlay (v0.2), and timeouts are marked as missed. Other features include `speak` for direct voice output (local playback on all platforms), `transcribe` for voice-to-text (supporting whisper-local/openai/macOS, with recording limited to macOS), and the `/voice` command to manage status. Voice synthesis uses local...
Read MorehuluTable: A backend-free online spreadsheet plugin for DeepSeek Harness Web
huluTable (npm package `dsh-hulutable-plugin`, MIT, v0.1) is a client-side plugin for DeepSeek Harness Web that provides a full-viewport table workspace in the sidebar.
Read Moreauth-gate: Adding Login Gate and Account Management to DeepSeek Harness Web Interface
`@dsh-login-gate/auth-gate` is a plugin bundle designed to provide access control and account management for the DeepSeek Harness Web interface. It is suitable for shared machines or LAN environments and requires no modifications to the DSH core. When not logged in, a fullscreen login page covers the entire application; after logging in, users can modify their username and password or log out via the "Account Settings" page within System Settings.
Read Moredsh-image-vision-bridge: Converts images to text descriptions for the DSH main model
# Introduction In agent development with DeepSeek Harness (DSH), a common frustration arises: the primary model is a pure text model, yet users upload screenshots to the conversation. DSH 0.1.1+ implements a built-in fallback for this scenario—requests do not error out, but the model can only see an image placeholder and cannot access the actual image content. To enable the model to truly respond based on images, one must either switch to a multimodal main model or implement a custom forwarding logic layer. dsh-image-vision-bridge solves this problem: it automatically forwards images from user messages to a vision model (defaulting to mimo-v2.5 on the opencode-go route), feeds the returned text description to the main text model, and preserves the original image display in the chat history. DSH's philosophy is "Everything...
Read Morebeav-creator-dsh: Integrating Xiaohongshu and Creator AI Workflows into DeepSeek Harness
beav-creator-dsh is a community plugin maintained by Jamailar (version 0.1.5, MIT-licensed connector) that directly integrates with the Harness tool via the Cordis service, without using MCP. It connects Beav's Xiaohongshu topic selection, trend research, knowledge management, copywriting, cover image generation, audio, short videos, and multi-step content production workflows to DeepSeek Harness. Long-running tasks are presented as native Harness jobs and replayable Beav task cards. Completion determination uses hard validation: a task is only accepted when the persistent artifacts reported by Beav are verified via readback through the Creator Gateway. Requirements include Beav 2.7.4+, DeepSeek Harness 0.1.0-rc.6+, and No...
Read Moredsh-theme-kit: 32 Theme Packs for DSH Web Interface
# Introduction I use the DSH Web interface daily for chatting and viewing outputs, but the interface's appearance is basically fixed. If you want to change the color scheme, the conventional approach is to edit CSS manually: finding selectors, overriding variables, and having to redo it all after every DSH upgrade. dsh-theme-kit turns this into a plugin. Once installed, you can change themes, wallpapers, adjust textures, and modify transparency directly within the settings panel, without writing any custom styles. DSH's philosophy is "everything is a plugin," so the appearance naturally falls to the plugin system. Below, I introduce this appearance kit maintained by ink5897 and released under the MIT license. # What is it dsh-theme-kit is an appearance plugin for the DeepSeek Harness Web interface. It is a...
Read Moredsh-web-terminal: Add a multi-tab local terminal to DSH web GUI
# Introduction When developing agents using DeepSeek Harness (DSH), most of the time is spent in the web GUI. However, whenever you need to execute shell commands—such as starting services, viewing logs, or temporarily entering vim to modify configurations—you have to switch to a separate terminal window and manually reconcile the directory and session state upon returning. `dsh-web-terminal` solves this problem: it adds a Terminal entry to the web GUI sidebar, opening a multi-tab terminal panel where each tab is a real PTY shell on the host process. DSH's philosophy is that "everything is a plugin"; therefore, this plugin requires no modification of DSH source code and can be installed directly into the web profile. The following sections cover the features, installation, configuration, and verification in order.
Read Moreopenspec-agents: Enable agent teams to implement OpenSpec changes following standardized processes.
When working with AI coding tools like DSH, the most common approach is to feed requirements directly into an AI and have it generate the entire codebase from scratch. While this is quick to start, the drawbacks are apparent: the process is opaque, and correctness relies entirely on post-hoc manual inspection; if an error occurs at any step, rolling back is difficult without digging through chat history; and starting a new session breaks the progress. OpenSpec (Fission-AI/OpenSpec, openspec.sh) addresses the first half of this workflow: writing specifications before code, organizing requirements into proposals, designs, tasks, and specs. However, what about after the specifications are written? Implementation, verification, review, and merging still require the user to manually oversee the AI step-by-step. openspec-agents fills precisely this missing gap:
Read Moredsh-rich-artifacts: A DSH plugin to publish workspace files as chat Artifacts
dsh-rich-artifacts is a DSH plugin developed by Inkotake (MIT licensed) designed to solve the inconvenience of users having to manually view files on the host machine after an agent delivers them. It publishes workspace files as chat Artifacts: the model can render PNG/JPEG/WebP/GIFs inline in the conversation by calling the `publish_artifact` tool (with parameters including path, title, and display mode), while other files are generated as cards with a download button. Files are stored in an ArtifactStore based on SHA-256 content addressing, with metadata stored in `~/.dsh/artifacts/v1/`. Persistent session events only record metadata without storing the file body, ensuring files remain accessible correctly after reloading or forking historical sessions.
Read MoreDSH GitHub Tool: Enabling Direct Connection of DeepSeek Harness to GitHub
The DSH GitHub Tool is an MIT-licensed open-source plugin (v1.0.0) maintained by H-table, enabling DeepSeek Harness (DSH) to connect directly to GitHub without involving the terminal. The plugin offers four automatically registered model tools: `github_api` (REST calls), `github_push_files` (single commit for multiple files, based on the git data API, supporting binary-safe local file upload), `github_graphql` (GraphQL calls), and `github_create_discussion` (create discussion posts). Tokens are encrypted and stored via the DSH credential system, not in configuration files; restricted networks can configure HTTP(S) proxies to connect directly. Install
Read Moredsh-llm-verifier: Verification, Ranking, and Controlled Application of Multi-candidate Patches
dsh-llm-verifier is a DeepSeek Harness plugin maintained by Web0926 that solidifies the workflow of "generating multiple candidate patches → test filtering → LLM ranking → manual approval and application". It runs 3 or 5 candidates for the same coding task, each working in an independent detached Git worktree. It first executes deterministic verification to eliminate failed candidates, then uses llm-verifier to rank them. If only one passes, it is selected directly; with 2 candidates, it uses a pivot; and with 3 to 5 candidates, it uses two pivots. The plugin imposes two approval gates: one before candidate execution and one before applying the patch. Before application, it verifies the repository path, base HEAD, clean status, and the patch's SHA-256. Verification commands can be automatically detected or explicitly specified and support npm/pnpm.
Read Moreshidi-skill: AI-for-Science Research Workflow Skill Pack for DSH
## Introduction When using Large Language Models (LLMs) for scientific research assistance, a common gap is that after a single round of conversation, the answers remain trapped in the chat history. Literature reviews cannot be archived, citation sources cannot be verified, and switching models requires re-explaining the context. For developers in the DeepSeek Harness (DSH) ecosystem, DSH's philosophy is "everything is a plugin." What is often missing is not model capability, but a reusable, verifiable research workflow that can be encoded into a skill. The shidi-skill introduced below is a skill package designed to address this issue: it fully executes tasks such as literature review, experimental design, plotting, and paper reading, delivering actual files to the disk. ## What is it shidi-skill is DeepSeek Harness's AI-for-Scien
Read Moredsh-vision-edge-doubao: Adding Vision Capability to DSH via Edge and Doubao Web
dsh-vision-edge-doubao is a DSH visual plugin developed by iceBear662 (MIT, v0.1.1), which drives the local Edge...
Read Moredsh-notebook-knowledge-studio: Run a NotebookLM-style local knowledge workbench in DSH Web UI
# Introduction Intelligent agent developers typically accumulate a wealth of materials locally: documents, scraped webpages, and pasted snippets. To use them for Q&A and organization, the common approach is to upload them to an online note-taking service and rely on its AI features—the trade-off being that the data leaves your local machine. If you are already using DeepSeek Harness (dsh) and agree with its "everything is a plugin" philosophy, you can actually mount a local-first knowledge workbench directly in the Web UI. The `iZY3/dsh-notebook-knowledge-studio` introduced below is exactly such a plugin: its workflow mimics NotebookLM, and data is stored in local Markdown files.
Read Moredsh-np-ppt: DSH Native PPT Presentation Expert Plugin
# Introduction Making agents produce a presentation that is ready for immediate delivery has always been a challenge in Agent development. The common approach is to have the model output Markdown or an outline, which is then manually laid out using templates; however, attempting to generate PPTX directly requires dealing with the complex OOXML format, which carries a high risk of errors. `dsh-np-ppt` integrates this into the DSH (DeepSeek Harness) plugin ecosystem: it uses a YAML-based PPTD DSL to describe slides, paired with a local visual editor and a Python-PPTX compilation kernel to complete the entire process from chat generation to PPTX export within a single pipeline. Below, we introduce the positioning, core capabilities, and installation and usage methods of this plugin. # What is dsh-np-ppt
Read Moredsh-md-table-formatter: Automatically aligns Markdown tables in assistant replies at the end of a turn.
dsh-md-table-formatter is a DeepSeek Harness plugin (v1.0.0, MIT) maintained by Hyna-hla. Based on the same surface-replace mechanism as the built-in compaction, it provides two functions: automatically aligning GFM tables in assistant replies at the end of a round, supporting three alignment markers, with widths calculated based on Unicode code points (CJK counts as 2 columns).
Read Moredsh-ponytail: Embedding the "Most Lazy Senior Engineer" mode into DeepSeek Harness
dsh-ponytail (npm package name: dsh-ponytail-skills) is a non-official port plugin that transplants the "laziest senior engineer" code style from DietrichGebert's ponytail into the Cordis plugin architecture of DeepSeek Harness (DSH), used to constrain the over-engineering tendencies of coding agents. The plugin registers six skills distributed with the package to the host layer of the `ctx.skills` registry: the ponytail core mode (prioritizing YAGNI -> reuse existing code -> standard library -> native capabilities -> one-liner -> minimal code, supporting lite/full/ultra intensity levels), code review targeting over-engineering, full repository audit, ponytail comment debt ledger, and receive...
Read Moredsh-llm-guardian: DeepSeek Harness Provider Health Check and Usage Statistics Plugin
# Introduction After configuring multiple Providers in DeepSeek Harness (DSH), several issues repeatedly arise: determining whether a Provider is currently available often requires waiting for a request to fail; checking account balance or subscription quotas requires opening a separate console page; and there is a lack of a unified view to see how many tokens were consumed locally over a period and on which models they were spent on. `dsh-llm-guardian` addresses these specific issues. It centralizes health checks, quota control, usage queries, and local statistics into the DSH settings page, while not saving API keys in plugin settings. Below is an introduction to what it does, how to install it, and what to watch out for when using it. # What is dsh-llm-guardian
Read More