DeepSeek Harness Plugin Hub

Publish and manage complete Harness Profiles. Discover Plugins for your next setup.

Explore

Plugins Presets Docs News

Community

Publish a plugin Contact Report an issue

Resources

Plugin Hub on GitHub DeepSeek Harness System status Privacy notice

© 2026 DeepSeek Harness Plugin HubPowered byPaxTech

Independent and unofficial. Not affiliated with, authorized by, or endorsed by DeepSeek.

CATEGORY

Security & Access DSH plugins

Verified manifests and exact versions in this category.

All categories

All plugins Agents & Orchestration1.6k Memory & Context2.1k Coding & Developer Tools4k UI & Customization5.6k Integrations & Communication3k Vision & Media919 Search & Research1.2k Security & Access1.8k Models & Usage2.5k Productivity & Workflow3.7k

All plugins

1808 plugins

@deepwhale-cn/dsh-env-audit

Env Audit

Has .env been committed, are its permissions correct, and is its content in plaintext?

Updated 10/16/2026

@deepwhale-cn/dsh-tool-vet

Tool Vet

See what a third-party plugin will affect before installing it

Updated 10/16/2026

dsh-phone-access

Phone Access

One-click mobile remote access in the DSH interface: automatically establishes a Cloudflare tunnel to securely expose the current session to a mobile browser, with the option to push the new address via 微信.

Updated 10/15/2026

@deepwhale-cn/dsh-consent-guard

Consent Guard

Check whether the consent mechanism is compliant: preselected by default, bundled consent, and no option to withdraw — the Personal Information Protection Law has clear requirements

Updated 10/15/2026

@deepwhale-cn/dsh-content-trust

Content Trust

Check whether the agent treats external content (web pages/files/emails) as instructions — an entry point for indirect prompt injection

Updated 10/15/2026

@deepwhale-cn/dsh-data-classify

Data Classify

Check whether data has classification labels and inventories — Article 21 of the Data Security Law requires establishing a classification and grading system

Updated 10/15/2026

@deepwhale-cn/dsh-data-rights

Data Rights

Check whether the rights to access, copy, and delete are actually implemented — Articles 45 and 47 of the Personal Information Protection Law explicitly require this.

Updated 10/15/2026

@deepwhale-cn/dsh-data-scan

Data Scan

Scan code and configuration files for ID numbers, bank card numbers, and phone numbers — with check digit validation to avoid false positives

Updated 10/15/2026

@deepwhale-cn/dsh-dep-audit

Dep Audit

Check the dependency tree for poisoning and risk exposure: install-time scripts, overly broad version ranges, and non-registry sources

Updated 10/15/2026

@deepwhale-cn/dsh-incident-package

Incident Package

If an incident is investigated, can you produce the evidence? Contingency plans, contacts, log preservation, reporting procedures—if you do not prepare in advance, it will be too late when trouble arises.

Updated 10/15/2026

@deepwhale-cn/dsh-key-rotate

Key Rotate

Check how long it has been since keys and certificates were changed, and whether there is a rotation policy—keys that are not changed for a long time often go unnoticed but pose real risks.

Updated 10/15/2026

@deepwhale-cn/dsh-least-privilege

Least Privilege

Check for excessive permissions: containers running as root, wildcard IAM, fully open CORS, tokens requesting admin — excessive permissions amplify intrusions

Updated 10/15/2026

@deepwhale-cn/dsh-license-guard

License Guard

Open-source license compliance check: scan the dependency tree to identify risks such as GPL/AGPL copyleft propagation and unknown licenses—run it before publishing, open-sourcing, or undergoing an audit.

Updated 10/15/2026

@deepwhale-cn/dsh-mcp-audit

Mcp Audit

Audit MCP configuration: unpinned remote packages, plaintext endpoints, executable local services — MCP is a new attack surface for agents

Updated 10/15/2026

@deepwhale-cn/dsh-pipia

Pipia

Check whether an impact assessment was conducted in accordance with Article 55 of the PIPL and retained for three years in accordance with Article 56 — processing sensitive information requires a prior assessment

Updated 10/15/2026

dsh-plugin-dm-security

Plugin Dm Security

Dameng DM8 database security assessment plugin: registers the dm_security_assess tool and performs 60 read-only checks across 13 domains (53 database-side and 7 host-side OS/SSH checks; host-side checks require additional SSH access). It provides six-state results (Critical/Warning/Notice/Compliant/

Updated 10/14/2026

Lock

Password gate: A password screen appears when opening DSH, and it can only be used after unlocking; supports manual and automatic locking, with passwords stored as one-way scrypt hashes.

Updated 10/14/2026

dsh-mcp-master

Mcp Master

MCP Master for DeepSeek Harness: global and project (.mcp.json) MCP servers, chosen per session from a picker beside the model selector — switchable at any time — with OAuth sign-in and trust review for project servers.

Updated 10/13/2026

dsh-strip-sandbox-permissions

Strip Sandbox Permissions

Strip placeholder tool-call arguments that DSH rejects: sandbox_permissions / justification (false sandbox escalation) and reasoning_effort (rejected subagent delegation).

Updated 10/13/2026

dsh-reasoning-bridge

Reasoning Bridge

Reasoning Bridge for dsh: hand high-cost reasoning to web AI (ChatGPT or Arena AI, multi-target), keep local evidence, adoption gate, edits, and tests in dsh. Contract-based Packet/Result, versioned risk consent, executable browser transport (zero-dependency CDP driver), SHA-256 run receipts.

Updated 10/13/2026

dsh-sovereign

Sovereign

DSH host-plane sovereignty clauses and content guard: injects sovereignty clauses into the final section of the system prompt, cleans credentials and covert payloads from tool results, adjudicates dangerous operations based on reversibility, and provides feature toggles that take effect immediately

Updated 10/11/2026

dsh-sandbox-mxc

Sandbox Mxc

DeepSeek Harness sandbox provider backed by MXC (Microsoft eXecution Containers): wraps ctx.sandbox.confine() argv as `wxc-exec --config-base64 <policy> -- <command>` on the BaseContainer backend. One runtime dependency (@microsoft/mxc-sdk, for the wxc-exec executable), no kernel patch.

Updated 10/11/2026

dsh-find-file

Find File

Permission-wall-immune file finder for DeepSeek Harness: registers the find_file agent tool. Unlike the built-in glob, partial failures never discard results — unreadable directories are skipped and reported, and complete:false warns that the tree was not fully covered. Zero runtime dependencies.

Updated 10/10/2026

dsh-cordis-mcp

Cordis Mcp

DeepSeek Harness 插件：把 DSH 的动态 Cordis 工具集（inspect/define/run/stop/undefine）以 MCP 暴露给 Claude Code。端点强制身份认证，token 可在 DSH 设置页配置。

Updated 10/10/2026

@eaglezlee/dsh-ui-html

Ui Html

Render sandboxed interactive HTML interfaces in DSH conversations, and send users' actions back to the model in structured form (by default as a new user message; in blocking mode as a tool result)

Updated 10/10/2026

dsh-peak-guard

Peak Guard

DeepSeek Harness 峰谷守卫：高峰计价时段直接拦下模型调用（一个字节都不发出去），谷时/周末/法定假期原样放行，并把拦下的请求记成一份可搜索、可勾选、可清理的列表 / A peak-pricing guard for DeepSeek Harness: during DeepSeek's peak pricing windows it short-circuits every model call before any bytes leave the machine, passes everything through off-peak, and keeps a searchab

Updated 10/9/2026

Pv0.1.0-alpha.29

@zeroclave/dsh-privacy

Privacy

ZeroClave privacy firewall for DeepSeek Harness

Updated 10/9/2026

@dsh-external/dsh-auto-reviewer

Auto Reviewer

Codex-style auto reviewer (approve for me) permission mode for DSH

Updated 10/9/2026

dsh-wsl-portproxy

Wsl Portproxy

WSL port expose advisor; local UI uses :3081/?token= on dsh ≥0.1.2.

Updated 10/9/2026

dsh-trustedrouter

Trustedrouter

TrustedRouter privacy and routing settings per model for DeepSeek Harness (dsh)

Updated 10/7/2026

← Previous 1…24 25 26 27 28…61 Next →

DeepSeek Harness Plugin Hub

Profiles Plugins Categories News Docs Sign in Manage Profiles

Profiles Plugins Categories News Docs Sign in