context safety score
A score of 40/100 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
cloaking
Page checks user-agent for bot/crawler strings to serve different content
cloaking
Page conditionally redirects based on referrer or user-agent
cloaking
Page loads content in transparent or zero-size iframe overlay
social engineering
The page uses aggressive urgency and social proof tactics: 'NEW: CLAUDE OPUS 4.6 LIVE NOW' repeated three times in a marquee banner, combined with inflated user claims ('Join +30M builders') to pressure users into signing up. These psychological manipulation techniques are characteristic of social engineering. (location: page-text.txt:9, page-hidden.txt:1-2)
brand impersonation
The site (blackbox.ai) explicitly co-opts the brand identities of Anthropic's Claude, OpenAI's Codex, and Google's Gemini by advertising 'Claude Code, Codex, Gemini, Blackbox. All Agents in One.' and 'Orchestrate Claude, Codex, Gemini & Blackbox agents from one interface.' This positions the site as an authoritative hub for all major AI platforms, which could mislead users into believing it is officially affiliated with or endorsed by Anthropic, OpenAI, or Google. (location: page.html:36, page-hidden.txt:5)
malicious redirect
A third-party analytics script is loaded from dubcdn.com with a data-domains attribute set to 'tryblackbox.link'. This redirects referral/attribution tracking through an alternate domain (tryblackbox.link), which is a common technique used to mask traffic sources, intercept referral data, or route users through intermediary domains for tracking or redirect purposes. (location: page.html:23)
hidden content
A third-party popup/overlay script is injected at the end of the body from 'smartarget.online' (https://smartarget.online/loader.js). This external script can inject arbitrary content, popups, or overlays into the page without being visible in the static HTML. The script source is not a recognized first-party or trusted CDN domain, and its behavior is opaque to the page operator and users. (location: page.html:427)
prompt injection
The page advertises a 'Chairman LLM' that selects agent outputs and supports 'Controllable Autonomy' with multi-agent orchestration across Claude, Codex, and Gemini. If users paste external content or code into this platform, adversarial content in that input could propagate prompt injection instructions to the underlying LLMs, especially given the multi-agent architecture described. This is a systemic risk inherent to the platform's design rather than explicit injected payload. (location: page-hidden.txt:5-6)
curl https://api.brin.sh/domain/blackbox.aiCommon questions teams ask before deciding whether to use this domain in agent workflows.
blackbox.ai currently scores 40/100 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.