nuclei for AI agents: the same findings, a fraction of the output. Part of Raijin.
-
Updated
Nov 19, 2026 - Go
nuclei for AI agents: the same findings, a fraction of the output. Part of Raijin.
AI-native offensive security tools. Signal, not noise.
Raijin by Atlantis helps teams build engineering culture by putting shared standards to work in their projects
Final Fantasy VIII Puzzle Game (Raijin Fishing). This was my application to demonstrate I knew enough to skip first semester of undergrad Computer Science in 2001. :D
katana for AI agents: every endpoint and parameter, collapsed by route pattern (/user/{id}), forms, APIs and exposed files ranked first. Part of Raijin.
katana for AI agents: endpoints, API routes and secrets found in JS bundles, deduped across files, ranked by severity. Part of Raijin.
ffuf for AI agents: only responses that differ from the baseline (status, size, words, timing), every input kept, soft-404s collapsed. Part of Raijin.
arjun for AI agents: every hidden parameter that changes the response, grouped by parameter set, SSRF/file/redirect/RCE/debug/IDOR names ranked first. Part of Raijin.
subfinder + dnsx for AI agents: every subdomain with its DNS answer, shared answers collapsed, dangling CNAMEs and internal IPs ranked first. Part of Raijin.
dalfox for AI agents: confirmed reflections with the payload and the context it landed in, one compact JSON. Part of Raijin.
CORS testing for AI agents: only exploitable policies (credentials + reflected, null or bypassed origin) as findings, each with a ready curl proof. Part of Raijin.
naabu + nmap for AI agents: every open port with its service, hosts with the same ports collapsed, exposed databases and admin services ranked first. Part of Raijin.
Subdomain takeover check for AI agents: dnsx plus a provider fingerprint table, dangling CNAMEs ranked first with the provider and the claim hint. Part of Raijin.
tlsx for AI agents: weak protocols/ciphers and certificate issues grouped by issue, every endpoint with its certificate and TLS config. Part of Raijin.
MCP server that exposes every Raijin tool to AI agents over stdio: one typed tool per binary, same compact JSON. Part of Raijin.
What changed between two runs of any Raijin tool: new, resolved and changed findings, every added or removed location, unchanged items counted. Part of Raijin.
httpx for AI agents: every URL with its status, title, tech and size, identical responses collapsed, the interesting hosts ranked first. Part of Raijin.
trufflehog for AI agents: verified secrets first, with the minimal context line, never the full file. Part of Raijin.
To associate your repository with the raijin topic, visit your repo's landing page and select "manage topics."