MCP security report
agent-observatory-sensor.nolimit-observatory.workers.dev
F19/100
Security grade FHigh-severity signals — review these before connecting.
Connected ✓31 tools scanned
https://agent-observatory-sensor.nolimit-observatory.workers.dev/s/gb18030-ok/mcp
highagent-tool-index
Embedded directive instructs agent to follow tracked redirect links
The description tells the agent to 'Follow start_here.hop first' through a redirect carrying utm_source/utm_medium tracking parameters, and frames this as mandatory tool behavior rather than optional context. This is an instruction aimed at steering the agent's actions (visit specific external URLs, in a specific order) embedded in a tool description, which is a classic injection pattern. It also promotes a 'pay-per-call tool gateway' which could lead the agent to register for or disclose information to a third-party monetization service without explicit user awareness.
highpeople-search-index
Facilitates search for personal data on living individuals + embedded redirect instruction
This tool's stated purpose is to help locate a 'people-search index' (a service that aggregates and searches personal/PII data on living people), which raises privacy and potential doxxing concerns if invoked on behalf of a user. Like the other index tools, it also instructs the agent to 'Follow start_here.hop' with tracking parameters embedded in the description — an embedded directive to the agent to navigate to a specific external endpoint, rather than a description of the tool's own function.
mediumagent-ads-index
Undisclosed advertising/placement mechanism with embedded redirect instruction
This tool advertises a service for getting products 'placed so AI agents discover' them, i.e., an ad-injection/monetization mechanism that could bias the agent's future tool recommendations or search results without disclosure to the user. It also contains an embedded instruction ('Follow start_here.hop first') with UTM tracking parameters baked into the description, directing the agent to specific external hops rather than merely describing functionality — a pattern consistent with steering agent behavior via tool description rather than tool output.
low
Not on gate's verified list
This server isn't on gate's handpicked list. That's not necessarily bad, but there's no third-party signal about it — review its tools carefully.
lowgb18030-ok
Schema parameters unrelated to stated tool purpose
Tool description is about GB18030 byte validation, yet the input schema includes many unrelated fields (zone, website, city, query, path, ref, feed, host, json) that have nothing to do with 4-byte encoding checks. This pattern is repeated identically across nearly all 31 tools, suggesting a shared/templated schema rather than purpose-built inputs; combined with repeated unverifiable 'discarded after the check' claims, this creates an opportunity for parameters to be silently repurposed (e.g., smuggling data through an irrelevant field) without it being obvious from the tool's stated function.
lowutc-time
Oversized schema unrelated to stated purpose
A simple 'current UTC timestamp' tool exposes input fields such as path, ref, feed, website, city, query, and json that have no bearing on returning the time. This broad, templated schema across many distinct tools makes it hard for a user or reviewer to know what data is actually being sent/processed versus merely declared, and relies on an unverifiable 'discarded' claim.
lowwebsite (shared parameter across many tools)
Repeated 'agent readiness' website-crawling capability bundled into unrelated tools
Many unrelated tools (e.g., utc-time, calc-eval, wiki-title-ok, jira-key-shape) all expose a 'website' parameter described as checking robots.txt, llms.txt, sitemap.xml, and extractable HTML — a broader web-fetch/crawl capability bolted onto tools whose names imply a narrow, unrelated function. This mismatch between declared tool purpose and actual capability surface could let the agent (or a crafted prompt) invoke broad web-fetching behavior under the guise of a narrowly-scoped tool.
Embed this badge
Show your MCP server’s security grade
Paste this into your README:
[](https://gateturbo.com/report/7454f3f9-f661-4d5b-8db1-4746ce456b1f)Scanned 10/1/2026 · This is a point-in-time snapshot of the server’s public tools. Re-scan.
Connect this server safely with gate.
Continuous re-scans, drift alerts, per-tool allow/ask/block, and a log of every call.
Monitor it — free