Jakob Wennberg
ec27228a8e
Em dashes (—) and en dashes (–) had spread across comments, docs, tests, and a few UI strings, reading as AI-generated boilerplate rather than house style. Replaced each with punctuation matching its context: colon for explanatory clauses, comma for asides, plain hyphen for numeric/legal ranges (e.g. "21-23§"), "to"/"till" for date ranges, parentheses for paired-dash asides. messages/en.json and messages/sv.json were fixed by hand together to keep sv/en in sync. Left untouched where the dash is the functional subject rather than decorative punctuation: date-range-parser.ts's separator regex, charset-repair.ts's CP1252 byte-mapping table (and its test), the SIE encoding mojibake docs, generic-csv.ts's minus-sign normalizer, the agent system-prompt files that already instruct against em dashes, and a golden iXBRL test fixture compared byte-for-byte. Also fixes two bugs surfaced along the way: an off-by-one in ApiKeysPanel's scope-label split (a leftover from an earlier partial pass), and a charset-repair test that had lost the literal en-dash it exists to verify. Regenerated the agent atom seed migration (skills:generate) since 27 SKILL.md files changed. Added a CLAUDE.md rule against em/en dashes, with an explicit carve-out for the functional-dash cases above. Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com>
60 lines
2.0 KiB
TypeScript
60 lines
2.0 KiB
TypeScript
/**
|
|
* Origin-header validation for the MCP Streamable HTTP endpoint.
|
|
*
|
|
* MCP spec (2025-06-18, Streamable HTTP transport): "Servers MUST validate
|
|
* the Origin header on all incoming connections to prevent DNS rebinding
|
|
* attacks." Also an explicit Claude Connectors Directory submission
|
|
* requirement.
|
|
*
|
|
* Non-browser clients send no Origin header and are allowed: claude.ai's
|
|
* backend connector, Claude Desktop, the npx gnubok-mcp bridge, Claude Code,
|
|
* and MCP Inspector (whose Node proxy makes the actual call). A browser page
|
|
* sends its own origin: allowed only when it matches the deployment's own
|
|
* host: compared against the request Host (covers Vercel previews and
|
|
* self-hosted domains without hardcoding) and NEXT_PUBLIC_APP_URL (covers
|
|
* proxies that rewrite Host). Anything else is a cross-site browser request
|
|
* the endpoint never serves (it sets no CORS headers), so reject explicitly.
|
|
*/
|
|
export function isForbiddenOrigin(request: Request): boolean {
|
|
const origin = request.headers.get('origin')
|
|
if (!origin) return false
|
|
|
|
let originHost: string
|
|
try {
|
|
originHost = new URL(origin).host
|
|
} catch {
|
|
// Malformed Origin (including the literal "null" some browsers send for
|
|
// sandboxed/opaque contexts): treat as foreign.
|
|
return true
|
|
}
|
|
|
|
const allowedHosts = new Set<string>()
|
|
const hostHeader = request.headers.get('host')
|
|
if (hostHeader) allowedHosts.add(hostHeader)
|
|
try {
|
|
allowedHosts.add(new URL(request.url).host)
|
|
} catch {
|
|
// request.url should always parse; ignore if not.
|
|
}
|
|
if (process.env.NEXT_PUBLIC_APP_URL) {
|
|
try {
|
|
allowedHosts.add(new URL(process.env.NEXT_PUBLIC_APP_URL).host)
|
|
} catch {
|
|
// Misconfigured env var: fall through to the request-derived hosts.
|
|
}
|
|
}
|
|
|
|
return !allowedHosts.has(originHost)
|
|
}
|
|
|
|
export function forbiddenOriginResponse(): Response {
|
|
return Response.json(
|
|
{
|
|
jsonrpc: '2.0',
|
|
id: null,
|
|
error: { code: -32600, message: 'Origin not allowed' },
|
|
},
|
|
{ status: 403 },
|
|
)
|
|
}
|