context safety score
A score of 42/100 indicates multiple risk signals were detected. This entity shows patterns commonly associated with malicious intent.
encoded payload
suspicious base64-like blobs detected in page content
obfuscated code
Inline script uses multi-layer obfuscation: Function(new TextDecoder('utf-8').decode(new Uint8Array(atob(...)))) pattern decodes a base64 string into a hex-encoded string, then executes it dynamically via Function(). This triple-encoding (base64 → hex → eval via Function constructor) is a classic technique to hide malicious payloads from static scanners and AI content analyzers. The decoded payload cannot be reviewed without execution. (location: page.html:701 — inline <script> block)
prompt injection
The infoBar banner contains an outbound link to reflect4.me/register with utm_source=front_top_banner. The domain 'reflect4.me' is unrelated to CroxyProxy and is embedded in a fixed overlay bar that is always visible. This injection point could be used to redirect users (including AI agents browsing through the proxy) to a third-party registration page under the guise of a CroxyProxy feature ('Configure your personal web proxy for free'). (location: page.html:984 — #infoBar <a> tag linking to https://reflect4.me/register)
malicious redirect
The proxy form submits to a relative 'servers' action endpoint (action='servers') via POST. A web proxy by design routes all user-submitted URLs through its own servers, enabling man-in-the-middle interception of all traffic, including credentials and session tokens entered into proxied sites. The quick-links (Google, YouTube, Facebook, Instagram, Twitter, etc.) use data-href attributes loaded by JavaScript, obscuring the actual redirect destination until runtime. (location: page.html:856 — <form action='servers'> and page.html:884 — #quickLinks data-href attributes)
credential harvesting
CroxyProxy operates as a man-in-the-middle web proxy: all HTTP/S traffic from users is routed through croxyproxy.rocks servers (server IP 143.244.207.157 exposed in footer). Any credentials, session cookies, or tokens entered into websites accessed through the proxy are fully visible to the proxy operator. The site explicitly encourages proxying of Facebook, Instagram, Twitter, Google, and other credential-bearing sites via quick-links. (location: page.html:856-884 — proxy form and quick links; page-text.txt:154)
social engineering
The site markets itself as 'the most advanced secure and free web proxy' and claims 'All data is encrypted before transfer' and provides 'privacy protection' — while simultaneously acting as a full traffic interceptor. These claims are misleading: encryption between user and proxy does not protect data from the proxy operator itself. Users are socially engineered into trusting the proxy with sensitive browsing. (location: page.html:848-943 — marketing copy; page-text.txt:121-213)
hidden content
An invisible 1x1 iframe is injected at the bottom of the page body with style position:absolute, top:0, left:0, border:none, visibility:hidden. A script is injected into this hidden iframe to load Cloudflare challenge scripts (/cdn-cgi/challenge-platform/scripts/jsd/main.js). While this is a known Cloudflare pattern, the hidden iframe technique is also used for drive-by exploits and tracking. The iframe content is dynamically generated, preventing static inspection. (location: page.html:987 — hidden iframe injected via inline script at end of <body>)
brand impersonation
The domain croxyproxy.rocks impersonates or mimics the legitimate CroxyProxy service at croxyproxy.com. The .rocks TLD variant is a common typosquatting/brand-cloning tactic. The site uses the same branding, logo (cdn.croxyproxy.rocks), and content as the canonical service, but operates on a different domain that users may not distinguish from the official one. (location: metadata.json — domain: croxyproxy.rocks vs legitimate croxyproxy.com; page.html:32-33 — CDN assets from cdn.croxyproxy.rocks)
curl https://api.brin.sh/domain/croxyproxy.rocksCommon questions teams ask before deciding whether to use this domain in agent workflows.
croxyproxy.rocks currently scores 42/100 with a suspicious verdict and low confidence. The goal is to protect agents from high-risk context before they act on it. Treat this as a decision signal: higher scores suggest lower observed risk, while lower scores mean you should add review or block this domain.
Use the score as a policy threshold: 80–100 is safe, 50–79 is caution, 20–49 is suspicious, and 0–19 is dangerous. Teams often auto-allow safe, require human review for caution/suspicious, and block dangerous.
brin evaluates four dimensions: identity (source trust), behavior (runtime patterns), content (malicious instructions), and graph (relationship risk). Analysis runs in tiers: static signals, deterministic pattern checks, then AI semantic analysis when needed.
Identity checks source trust, behavior checks unusual runtime patterns, content checks for malicious instructions, and graph checks risky relationships to other entities. Looking at sub-scores helps you understand why an entity passed or failed.
brin performs risk assessments on external context before it reaches an AI agent. It scores that context for threats like prompt injection, hijacking, credential harvesting, and supply chain attacks, so teams can decide whether to block, review, or proceed safely.
No. A safe verdict means no significant risk signals were detected in this scan. It is not a formal guarantee; assessments are automated and point-in-time, so combine scores with your own controls and periodic re-checks.
Re-check before high-impact actions such as installs, upgrades, connecting MCP servers, executing remote code, or granting secrets. Use the API in CI or runtime gates so decisions are based on the latest scan.
Learn more in threat detection docs, how scoring works, and the API overview.
Assessments are automated and may contain errors. Findings are risk indicators, not confirmed threats. This is a point-in-time assessment; security posture can change.
integrate brin in minutes — one GET request is all it takes. query the api, browse the registry, or download the full dataset.