context safety score
A score of 36/100 indicates multiple risk signals were detected. This entity shows patterns commonly associated with malicious intent.
hidden instruction
high hidden content ratio detected in DOM
phishing
The domain zebra.engineering is serving a pixel-perfect clone of the Google Sign-In page (accounts.google.com). The HTML sets <base href="https://accounts.google.com/v3/signin/"> to mask the true origin, presents Google branding, and collects email/phone credentials via a standard Google sign-in form flow. The actual serving domain is zebra.engineering, not accounts.google.com. (location: page.html:1 — <base href="https://accounts.google.com/v3/signin/">)
brand impersonation
The page fully impersonates Google's sign-in UI including Google's CSS/JS assets from gstatic.com, Google Sans fonts, Google color scheme (#0b57d0), Google logo markup, and the text 'Sign in with Google'. The page is hosted on zebra.engineering, a completely unrelated domain with no affiliation to Google. (location: page.html:1 — full page structure, base href, gstatic.com asset references)
credential harvesting
The fake Google sign-in form collects email or phone number inputs. The form action and OAuth flow parameters (client_id=995294672929-548avbaagkutnug1m5tuprl85chphgad, redirect_uri=https://iap.googleapis.com/...) are present but the page is served from zebra.engineering, meaning submitted credentials are exposed to the attacker-controlled server before or instead of Google. (location: page.html:29 — form with identifier input, OAuth parameters including client_id and redirect_uri)
malicious redirect
The page embeds a complex OAuth redirect chain with a continue parameter pointing back to accounts.google.com/signin/oauth/legacy/consent, creating a believable redirect flow that masks the phishing origin. The base href override combined with relative URL resolution routes all form submissions and link clicks through accounts.google.com in appearance only. (location: page.html:29 — href containing continue=https://accounts.google.com/signin/oauth/legacy/consent and redirect_uri=https://iap.googleapis.com/...)
obfuscated code
The page contains heavily obfuscated JavaScript using opaque state-machine control flow (while loops with numeric state variables like p=69, M=66), bitwise operations, and single-letter variable names throughout. Function F contains an iframe-based exfiltration mechanism that dynamically creates iframes, monitors their load/error events, and collects timing telemetry — consistent with credential-capture relay or bot-detection bypass infrastructure. (location: page.html:16 — function F with iframe creation loop; page-text.txt:2 — obfuscated JS block)
prompt injection
The page-text.txt visible text contains raw JavaScript source code inline (including sourceMappingURL with base64-encoded data URI and large obfuscated script blocks) that would be ingested verbatim by any AI agent scraping or summarizing page text. An AI agent processing this page's text content would receive executable-looking instructions and obfuscated code as apparent 'content', which could confuse agent reasoning or trigger unintended code interpretation. (location: page-text.txt:1 — sourceMappingURL data:application/json;base64,... embedded in visible text; page-text.txt:2 — raw JS injected into text layer)
hidden content
The <base href> tag redirects all relative URLs to accounts.google.com, effectively hiding the true origin of all resource loads and form submissions from casual inspection. Additionally, the botguard integration (window.botguard.bg) and hiddenMultipleChoiceIdentifier element are hidden form fields used to silently capture bot-detection signals alongside user credentials. (location: page.html:1 — <base href="https://accounts.google.com/v3/signin/">; page.html:27-28 — hiddenMultipleChoiceIdentifier, bgresponse hidden inputs, botguard invocation)
curl https://api.brin.sh/domain/zebra.engineeringCommon questions teams ask before deciding whether to use this domain in agent workflows.
zebra.engineering currently scores 36/100 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.