dsh-session-health: Read-Only Health Diagnostics for DSH Multi-Frame Zstd Session Files
dsh-session-health is a DSH admin-security plugin maintained by omdsh-dev. It performs frame-level read-only scans on multi-frame zstd session files located under $DSH_HOME/sessions, detecting torn, corrupted, empty sessions, and stray files. The plugin registers the `session_health` tool, which outputs JSON reports and cleanup recommendations. It supports three actions: scan, file, and stats, with an optional deep analysis mode. It can be installed via `dsh plugin add` in DSH 0.1.0-rc.8 and complements the dsh-session-repair-skill. Licensed under MIT.
Read Moredsh-auto-approve: Adding Auto Permission Level Between Workspace Write and Full Access
dsh-auto-approve inserts an auto permission tier between workspace-write and danger-full-access for DeepSeek Harness. The sandbox boundary is identical to workspace-write, with routine sandbox escalations approved in one go by a classification model; cases hitting a deterministic danger list, where the classifier is uncertain, or due to timeout or malformed responses are still escalated to manual approval. The plugin supports Web, TUI, and Desktop frontends and can be installed via dsh plugin and enabled by activating the Auto tier in Permissions.
Read More@awiki/dsh-plugin: Integrating AWiki Identity and Messaging in DeepSeek Harness
Introduction to the DSH plugin @awiki/dsh-plugin maintained by AgentConnect (SkillHub: agentconnect/dsh-awiki). It covers AWiki identity registration and Recovery V4 restoration, web-based private/group chats, agent messaging and email tools, optional listener and standalone Model Proxy package, and provides Profile installation commands, key environment variables, and usage notes. The plugin is MIT-licensed, while the Rust IM Core dependency is AGPL-3.0-only.
Read MoreRisk Rule Design: Full Workflow of Ventilation Control Rule Mining in DSH
risk-rule-design is a third-party plugin released by pypcfx-glitch for DeepSeek Harness, distilled from the "100-Day Risk Control Expert" methodology. It performs data quality checks on samples containing X/y, mines single rules, finds the optimal combination of parallel rule sets, and outputs a self-contained HTML analysis report. The plugin registers three tools: rrd_profiling, rrd_mining, and rrd_report, along with the risk-rule-design expert skill. It is pure JavaScript with zero dependencies and is open-source under the MIT license. It is recommended to install it via `dsh plugin --profile web install dsh-plugin-risk-rule-design`, and invoking the skill in a conversation will seamlessly integrate the entire workflow.
Read Moredsh-mcp-manager: A MCP Management Plugin for DSH with OAuth and stdio Capabilities
dsh-mcp-manager is a DSH web plugin maintained by hyqhyq3, managing MCP servers on the Settings → MCP page. It supports HTTP OAuth (PKCE + dynamic client registration), static tokens (environment variable names not stored on disk), and stdio local processes, with tools registered under the mcp__<name>__* namespace. Optional features include on-demand proxy schema compression, support for workspace-level mcp.json, and exclude blocking. Installation: npx -p @deepseek-ai/dsh dsh plugin --profile web add github:hyqhyq3/dsh-mcp-manager.
Read Moredsh-web-billing: RMB/USD Token Billing Plugin for DSH Web
dsh-web-billing is a DeepSeek Harness Web-based billing plugin maintained by bpc-oss, which automatically calculates pricing according to official policies (including peak and off-peak pricing as of 2026-08-17). It supports coding plan routing by provider, message-level ledger tracking, and a summary of account balance and settings page fees. After installation, restart the DSH web and navigate to Settings → Fees to view expenses and export billing statements. Local statistics are unofficial; please review the source code and MIT license before installation.
Read Moredsh-approve-for-me: Rule-Driven Sandbox Automatic Approval
dsh-approve-for-me is a DeepSeek Harness plugin maintained by Timeance, designed for rule-driven automatic approval of privilege escalation in Shell and PowerShell sandboxes. It sequentially performs fixed high-risk checks, literal command prefix rules, and optional LLM reviews, granting only a single "allowed-once" upon approval and deferring to the Harness's native manual approval when uncertain. This article introduces the installation command, Access preset activation methods, Web/YAML configuration examples, and the security model, catering to DSH users who aim to reduce repetitive approvals without switching to Full access.
Read MoreDSH Config Manager: One-Click Backup, Migration, and Synchronization for DSH Configurations
dsh-config-manager is an admin-security configuration management plugin for DSH, which supports exporting and importing settings, plugins, MCP, Skills, workspaces, etc., as ZIP files. It offers scheduled backups, Git/WebDAV remote synchronization, a configuration marketplace, and snapshot rollback. By default, keys are not exported, and before importing, users can preview and handle path mapping and conflicts. Installation command: dsh plugin --profile web add dsh-config-manager@latest.
Read Moredsh-security-audit: DSH Native Security Audit and Data Desensitization Risk Report
dsh-security-audit is a native security audit plugin (admin-security) for DSH, maintained by omdsh-dev. It scans configuration and credential metadata, plugin sources, session structures, and network exposure surfaces in a read-only manner, outputting desensitized and reproducible JSON risk reports. It provides six actions: scan_config, scan_plugins, scan_sessions, scan_network, report, and rules. It does not perform automatic repairs, connect to the internet, or execute audited plugins. It is compatible with DSH 0.1.0-rc.8 and can be installed via `dsh plugin add` from GitHub or npm pack.
Read Moredsh-plugin-healthcheck: Install Plugin, Verify First, Then Go Live
dsh-plugin-healthcheck is a DSH plugin health check tool maintained by chenw2759-wq, categorized under admin-security. It provides a "Plugin Inspection" feature in the web settings panel, integrating L0 static checks (file integrity, dependency declarations, high-risk duplicates, cordis usage, etc.), L1 configuration combinations, L2 isolated trial runs (complete boot of child processes without restarting the backend), and C8 pure static trojan scanning. When issues are detected, it supports one-click fixes, automatic rollback (writing home patch to disable lines, HMR hot-reload), or copying predefined prompts for agent use. The ironclad rule for fixes restricts modifications to plugin code and configuration layers only, strictly prohibiting changes to harness source code. Licensed under MIT, with 14 stars on GitHub.
Read Moredsh_plugin_swift_cycle: DSH Adapter for Swift Cycle Governance Skill
dsh_plugin_swift_cycle is the DeepSeek Harness adapter for Swift Cycle v1.2.0, maintained by Solismuchengxue and categorized under admin-security. It registers governance capabilities such as Knowledge Promotion, State Separation, and Governance Baseline as Skills (/swift-cycle) that are only explicitly invoked by users, with modelInvocable set to false. The adapter version is 0.1.1, supporting installation via npm or pinned GitHub commits, and payloads can be verified offline. It is suitable for project teams that require version reproducibility and clear governance boundaries; please check the source code and MIT license before installation.
Read Moredsh-sandbox-escalation-fix: Projecting Sandbox Upgrade Schema Per Session to Fix Third-Party Model Tool Invocation Loop
dsh-sandbox-escalation-fix is a DSH community plugin maintained by JUSTMONIKA2022, providing a zero-configuration fix for third-party models like GPT in All Access mode. It addresses bash, pwsh, write, and edit validation failures and retry loops caused by static sandbox schema upgrades. The plugin dynamically projects model-visible parameters based on the Session's Sandbox Mode and Approval Policy, synchronizes Native Tool Call with the Code Mode SDK, and only processes redundant upgrade requests of the same mode during execution. It supports DSH versions 0.1.0-rc.5 through rc.8, as well as 0.1.1-rc.1/rc.2, and is released under the MIT license.
Read Moredsh-secure-audit: DSH Read-Only Security and Compliance Audit Plugin
dsh-secure-audit is a read-only security plugin (admin-security) for DeepSeek Harness maintained by PensiveFei. It provides security_scan_text for prompt injection detection, security_redact_text/security_redact_json for PII and JSON redaction, security_audit for local configuration auditing, and security-review skill. The plugin does not write to or execute the audited system; reports are redacted and include reportSha256 self-verification. Installation: dsh plugin add dsh-secure-audit; requires Node ≥20 and @deepseek-ai/dsh-tools >=0.1.0-rc.7.
Read Moredsh-clawrouter: Adding a Dangerous Tool Call Review Gate to DeepSeek Harness
dsh-clawrouter is a DSH plugin maintained by BlockRunAI that audits and forcibly intercepts dangerous tool calls before execution using a more powerful model, while also enabling access to 67 models via BlockRun routing on a pay-per-use basis with a single wallet. This article introduces the review gate, BlockRun routing, installation commands and cordis.patch.yml configuration, as well as the usage and precautions of built-in commands like /gate and /review.
Read Moredsh-full-remote: An Auditable Remote Access Gateway for DeepSeek Harness
dsh-full-remote is a DeepSeek Harness plugin maintained by JUANWANG-BUAA. It inserts a token-authenticated reverse proxy in front of the Harness Web service, rewrites the Host/Origin to loopback addresses, and prevents privileged APIs such as settings, credentials, and host.listDirectory from returning 403 errors during public network tunneling or local area network remote access. It supports device sessions, mobile QR code invitations, JSONL audit logs, and optional Cloudflare Quick Tunnel. The installation command is: dsh plugin --profile web add dsh-full-remote.
Read Moredsh-custom-tool: Creating and Managing Sandbox JavaScript Tools in DSH
dsh-custom-tool is a DSH plugin (admin-security, MIT, v0.1.2) maintained by omdsh-dev. Users can write JavaScript tools using the Monaco editor in the setup UI, and models can self-service add or delete tools through APIs like custom_tool_create. Tools are persisted in the custom-tools namespace and hot-registered, executing within node:vm workers with global/workspace scope and permission boundaries. This article introduces installation commands, code contracts, configuration options, and security considerations.
Read Moredsh-usage-stats: DSH Usage Heatmap and Token Dashboard
dsh-usage-stats is a DSH community plugin maintained by Make0209. It aggregates usage data from persistent session logs, providing a GitHub-style 53-week heatmap, token/cache hit/account balance statistics cards, and workspace alias management. The installation command is `dsh plugin --profile web add dsh-usage-stats`, which takes effect upon refresh and automatically backfills historical data. It is ideal for developers who need to track local tokens and continuous usage by workspace.
Read Moredsh-computer-use: Foreground-Safe Native macOS Control Capabilities for DSH Agent
dsh-computer-use is a DSH plugin maintained by Anionex, providing Accessibility-first computer control capabilities for macOS 14+. It is centered around fresh observation, expired state rejection, and scoped permissions. By default, it does not move the system cursor or perform global pointer injection, instead dispatching fallback input to target processes and windows. The article introduces the plugin's positioning, core functions, npm installation commands, Skill loading methods, and typical workflows, while also explaining applicable scenarios, permission requirements, and security considerations.
Read Moredsh-plugin-guard: DeepSeek Harness Plugin Installation Safety Net
dsh-plugin-guard is a DSH admin-security plugin maintained by lxzy-7. It automatically snapshots configuration files before each plugin change and implements automatic rollback through boot-guard daemon startup and health checks. In case of incidents, it generates reports to trigger Agent analysis. It supports Web UI backup management, dsh-guard CLI, and Agent tools, and starting from v0.3.2, it can isolate problematic plugins during rollback failures. It is suitable for DSH users who frequently install community plugins, require reversible changes, and need self-healing for startup failures.
Read Moredsh-remote: Adding a Remote Access Authentication Layer for DeepSeek Harness
Add account password login, TOTP MFA, signed session cookies, and role-based permissions before the DSH web interface. Replace workspace selection with an in-browser directory dialog. After authentication, allow DSH's originally loopback-only privileged APIs to be accessed. Suitable for scenarios where DSH is exposed to a local network or the internet via a reverse proxy or tunnel. Install with the command: dsh plugin --profile web add @xgone/dsh-remote. MIT license, maintained by xgone.
Read Moredsh-claude-ux: Replicating Claude-style Area Risk Control and Dialog Autonomy for DSH Web
dsh-claude-ux is a DSH web profile plugin maintained by eri64, replicating Claude's regional risk control (reversible) and conversational autonomy: local signal determination, rejection ladder, lexicon plus LLM abuse adjudication, and proactive termination for severely harmful content. It does not communicate externally by default, is installed via a single npx command, configured through the "Claude Risk Control" tab in the settings page, and is licensed under MIT.
Read Moredsh-reverse-skill: Encapsulating reverse-skill completely as a DSH Cordis plugin
dhicoc/dsh-reverse-skill packages the 87 SKILL.md files from the upstream reverse-skill as-is into a DeepSeek Harness Cordis plugin, covering reverse engineering, authorized penetration testing, and security research. Compared to the preset solution that only ports 45 skills, this plugin automatically registers all skills via ctx.skills, enabling activation with a single "dsh plugin add" command and supporting fallback mounting through customSkillDirs.
Read MoreDockyard DSH: Integrating Multi-Provider Account Pools for DeepSeek Harness
Dockyard DSH is a native DSH plugin maintained by AITabby that integrates official OAuth services from Codex, Antigravity, Grok, Claude, and Cursor, as well as local sessions, into a unified account pool. It provides the `/dockyard` command to manage login, scanning, policies, and real-time model quotas without requiring an additional proxy gateway. A DMG version is available for macOS, and it can also be installed via `dsh plugin --profile web add`. This article introduces its positioning, core capabilities, installation steps, and typical commands, while also explaining security boundaries and platform limitations.
Read MoreDeepSeek Harness Auto-Approval: dsh-auto-mode Bridges the Middle Layer in Permission Modes
Introducing the DSH plugin dsh-auto-mode (@nanmicoder/dsh-auto-mode) maintained by NanmiCoder. It adds an Auto/automatic approval mode on top of the official workspace-write sandbox: routine builds, tests, and dependency installations are automatically allowed, while semantic risks outside the sandbox are reviewed by a classifier, and dangerous operations such as accessing the root directory or credential paths are directly rejected. The article covers a comparison of four permission modes, installation commands, steps to enable it in the Web UI, optional classifier configuration, and an explanation of security boundaries.
Read More