hunt-cors · git:20260929.29bbcb8 · 2026-09-29 · sha256 e309f044ec17610a
hunt-cors git:20260929.29bbcb8A
Immutable. This exact content never changes and is served at /api/v1/blob/e309f044ec17610a.
No tool to install first
claude plugin marketplace add https://markdownregistry.com/mp/a/art_vaurjads7tmnpc2y/e309f044ec17610a93534617201a579d64788b4ac4fa6c5a309a903dc67f602e.json
Then /plugin install hunt-cors@markdownregistry-vaurjads7tmnpc2y in Claude Code. It installs this exact version, the one graded A here, pinned to the commit it was graded at, and leaves it there.
--- name: hunt-cors description: CORS — origin reflection, null origin, preflight abuse, wildcard detection. Establishes whether credentialed cross-origin reads are actually possible. --- # CORS ASSESSMENT **Core:** the only question that matters is whether a hostile origin can read an authenticated response. A reflected `Origin` without `Access-Control- Allow-Credentials: true` is not exploitable and must not be reported. ## DETECTION ```bash curl -s -I -H "Origin: https://evil.com" "https://target.com/api/userinfo" | grep -i "access-control" # ACAO reflects evil.com AND ACAC: true → credentialed read is possible curl -s -I -H "Origin: null" "https://target.com/api/userinfo" | grep -i "access-control" # Also check the OPTIONS preflight response, which is what the browser enforces curl -s -X OPTIONS -H "Origin: https://evil.com" \ -H "Access-Control-Request-Method: POST" \ -H "Access-Control-Request-Headers: authorization" \ "https://target.com/api/userinfo" -D- -o /dev/null ``` ## WHAT ACTUALLY MATTERS ``` Origin reflected + ACAC: true + sensitive response → exploitable Origin reflected + ACAC absent/false → NOT exploitable ACAO: * (no credentials) → NOT exploitable on authed data Null origin + ACAC: true → exploitable from a sandboxed iframe Trusted subdomain list + wildcard → check which hosts are actually trusted HTTP endpoint missing origin check → may bypass an HTTPS-only policy ``` ## PROOF (minimum sufficient) To confirm the read, show the response body actually crosses origins under `withCredentials` — a PoC page plus the returned payload. **Do not then use the stolen data to act** (no account takeover, no state change). A single demonstrated cross-origin read is the complete finding. ## REPORTING NOTES - Report the credentialed read and the data it exposes. That is the severity. - Where it *could* compound with XSS on a trusted subdomain, note it as a compound-risk line; do not go find the XSS to justify a higher severity. - `Access-Control-Allow-Origin: *` on public, non-credentialed endpoints is correct behavior — reject it (Gate 6).