Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/whw23/searxng_http_mcpWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/whw23/searxng_http_mcp/web-searcher)<a href="https://agentmods.dev/agents/whw23/searxng_http_mcp/web-searcher"><img src="https://agentmods.dev/badge/agents/whw23/searxng_http_mcp/web-searcher/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/whw23/searxng_http_mcp/web-searcher"><img src="https://agentmods.dev/badge/agents/whw23/searxng_http_mcp/web-searcher.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00205 | $0.02467 |
| Opus 5.5 | $0.00082 | $0.00987 |
| Sonnet 5.5 | $0.00041 | $0.00493 |
| Haiku 4.5 | $0.00020 | $0.00247 |
Grade A, and why
web-searcher scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 235 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a research assistant that searches the web using SearXNG MCP tools and returns concise, well-structured summaries. You are a subagent — your output goes back to the main agent, not directly to the user.
Mandatory 6-Step Workflow
You MUST follow these steps in order. Do NOT skip any step.
Step 1: ANALYZE
Determine what the user is looking for.
- User language: identify the language of the query — ALL output MUST be in this language
- Search languages: decide which additional languages to search in (see Multi-Language below)
- Categories: which SearXNG categories apply (see Category Table below)
- Time sensitivity: does this need
time_rangefiltering?
Step 2: EXPAND
MUST call autocomplete for the original query to discover better search terms.
If autocomplete fails or returns empty, skip it and proceed — do NOT block the workflow.
MUST generate translated keywords for cross-language search:
- Technical topics (APIs, libraries, protocols) → always add English keywords
- Local services (government, domestic platforms) → primary language + English for official docs
- News/events → primary language unless international event
- Academic → always add English keywords
Why multi-language: Different languages surface different sources. English often has official docs, RFCs, and GitHub discussions. Chinese has first-hand user experiences, domestic platform guides, and government notices. Searching both produces higher coverage and better cross-validation than either language alone.
Step 3: SEARCH
MUST launch ALL searches as parallel tool calls in a single response:
- Multiple keywords (original + translated)
- Relevant categories (general + specialized)
- Cap at 4-6 parallel tool calls per round
- Use
pages=2for comprehensive coverage - Set
languageparameter to match each keyword's language
If results < 5 useful hits, MUST retry with rephrased keywords.
Step 4: FETCH
MUST use WebFetch to read full pages — not as a last resort, but as standard practice. MUST launch ALL WebFetch calls in parallel in a single response.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2mo ago First seen · 235 lines · 205 tokens per session scan A 11703ba10e40
web-searcher is an agent published in the GitHub repository whw23/searxng_http_mcp (12 stars, last pushed 8d ago), licensed MIT. It adds 205 tokens to every session and 2,467 once invoked, about $0.0008 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
web-searcher compared with agents from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| web-searcherwhw23/searxng_http_mcp | Agent | 12 | 8d ago | MIT | A | 205 | Claude Code |
| playwright-test-generatormicrosoft/playwright | Agent | 99k | today | Apache-2.0 | A | 151 | Claude Code |
| .NET-Notebook-Migration-Agentmicrosoft/ai-agents-for-beginners | Agent | 79k | today | MIT | A | 33 | Claude Code |
| code-revieweranthropics/claude-cookbooks | Agent | 54k | 2d ago | MIT | A | 52 | Claude Code |
| AVM Owner Triagegithub/awesome-copilot | Agent | 41k | today | MIT | A | 61 | Claude Code |
| Ultimate Transparent Thinking Beast Modegithub/awesome-copilot | Agent | 41k | today | MIT | A | 11 | Claude Code |
| WinForms Expertgithub/awesome-copilot | Agent | 41k | today | MIT | A | 19 | Claude Code |