dsh-web-search
Run the following command in DeepSeek Harness:
dsh plugin install lemoncat7/dsh-web-search
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install lemoncat7/dsh-web-search in the DeepSeek Harness terminal to install the web search plugin from https://github.com/lemoncat7/dsh-web-search.
About this plugin
DeepSeek Harness agents that need to search the web or read public pages often lack a safe, governed channel. dsh-web-search fills that gap by consolidating web_search, web_fetch, and a constrained web_source tool into a single plugin-scoped boundary. Administrators pick a backend, store credentials, and test the draft in the settings panel — all without restarting DSH.
The core design centers on security boundaries. Five search backends are supported: self-hosted SearXNG, keyless Chinese Wikipedia, agent-oriented Tavily, traditional Brave Search for web and news, and Gemini Grounded Search with URL context. Each query is dispatched to exactly one selected provider; there is no silent fan-out or fallback quota spending. On the fetch side, only credential-free public HTTP(S) URLs are accepted. Loopback, link-local, private, and reserved destinations are rejected outright. Redirects are capped at five hops with revalidation at every step, binary bodies are discarded, and response payloads are limited to 2 MiB. web_source further restricts access to bounded range reads or a bounded set of exact source markers, capping output at 120,000 characters. API keys are resolved through the DSH Credential Store and appear only in provider-defined authorization headers; browser settings routes enforce exact same-origin writes and never return secret values.
Who is it for? If your team needs agents on the web but demands strict boundaries — especially if you prefer a self-hosted SearXNG instance, want to switch among multiple search backends, or require proxy and credential isolation to stay within a single plugin scope — dsh-web-search delivers a complete, security-first path from configuration to runtime.
Use Cases
- Agents need web research while administrators enforce a single search backend with isolated credentials and security boundaries
- Teams self-host SearXNG as the search backend and hot-swap to other APIs without restarting DeepSeek Harness
- Agents must verify raw HTML markers or script-embedded data with bounded, read-only range reads under a constrained source procedure
Best For
- Development teams building agent applications with DeepSeek Harness that require web access capabilities
- Security or ops teams enforcing strict credentials management, intranet isolation, and response-size limits
- Advanced users preferring self-hosted search services with flexible multi-backend switching
Related Plugins
Free web search plugin for DeepSeek Harness with web search, X search, and page fetch; no signup or API key required, with automatic multi-engine failover.
AnySearch-powered web and vertical search plugin for DeepSeek Harness, offering real-time search, cleaned URL content, concurrent batch search via native web_search/web_fetch, no API key required.
Pixel-perfect webpage clone tool that uses an agent harness to turn any webpage into a scored, full-page React replica.
A bilingual cost-tracking plugin for DeepSeek Harness with session/daily cost, budget, official & custom provider balance, coding plan quotas, peak/off-peak pricing alerts, and history stats.