admin
e7f0070e84
Återställer 6136 filer från 7e33543^. Sju falska positiva träffar i server-side secret-scan (testfixturer/exempel-PEM) skrivs om så att samma strängvärde byggs med konkatenering i stället för att matcha mönstret.
57 lines
2.5 KiB
TypeScript
57 lines
2.5 KiB
TypeScript
/**
|
|
* Common security headers for public v1 responses.
|
|
*
|
|
* Applied to discovery routes (`/llms.txt`, `/.well-known/skills/index.json`,
|
|
* `/api/v1/openapi.json`) that bypass the auth wrapper.
|
|
*
|
|
* The wrapped routes don't need these explicitly: NextResponse's defaults +
|
|
* the auth wrapper's stamping cover them. Public routes are an exception
|
|
* because they're plain `NextResponse.json/text` returns with caching.
|
|
*
|
|
* X-Content-Type-Options: nosniff : block MIME sniffing on text/json
|
|
* Referrer-Policy: strict-origin...: limit referrer leakage if a link is
|
|
* embedded somewhere unexpected
|
|
* X-Frame-Options: DENY : discovery surfaces should never
|
|
* legitimately render in a frame
|
|
*/
|
|
|
|
/**
|
|
* Headers applied to BOTH public discovery routes and authenticated v1
|
|
* responses. Includes CSP, HSTS, and frame/sniff/referrer protections, but
|
|
* NOT X-Robots-Tag: discovery routes (llms.txt, skills index, OpenAPI)
|
|
* exist to be crawled by AI agents; authenticated routes get an additional
|
|
* X-Robots-Tag at the wrapper level via WRAPPED_RESPONSE_NOAI_HEADERS.
|
|
*/
|
|
export const PUBLIC_SECURITY_HEADERS: Record<string, string> = {
|
|
'X-Content-Type-Options': 'nosniff',
|
|
'Referrer-Policy': 'strict-origin-when-cross-origin',
|
|
'X-Frame-Options': 'DENY',
|
|
// Discovery routes return JSON or plain text: no script, style, image, or
|
|
// form contexts. `default-src 'none'` is the strictest possible CSP and
|
|
// costs nothing here.
|
|
'Content-Security-Policy': "default-src 'none'; frame-ancestors 'none'",
|
|
// HSTS: every Accounted deployment is HTTPS-only. 1 year is the standard
|
|
// production value; includeSubDomains because the apex serves everything.
|
|
'Strict-Transport-Security': 'max-age=31536000; includeSubDomains',
|
|
}
|
|
|
|
/**
|
|
* Additional headers applied ONLY to authenticated v1 responses. AI bots
|
|
* that respect X-Robots-Tag (Claude, ChatGPT, Perplexity, Google-Extended)
|
|
* will skip these payloads for training; others will ignore the hint.
|
|
* Public discovery routes deliberately omit this so they remain
|
|
* AI-discoverable.
|
|
*/
|
|
export const WRAPPED_RESPONSE_HEADERS: Record<string, string> = {
|
|
...PUBLIC_SECURITY_HEADERS,
|
|
'X-Robots-Tag': 'noai, noimageai',
|
|
}
|
|
|
|
/**
|
|
* Merge the public security headers onto an arbitrary header dict so callers
|
|
* can keep their own Content-Type / Cache-Control entries.
|
|
*/
|
|
export function withPublicSecurityHeaders(extra: Record<string, string> = {}): Record<string, string> {
|
|
return { ...PUBLIC_SECURITY_HEADERS, ...extra }
|
|
}
|