docs(backend): pin get_public_key as an intentional, load-bearing ACL exception (#26) #53
1 changed files with 23 additions and 0 deletions
docs(backend): pin get_public_key as an intentional, load-bearing ACL exception (#26)
Some checks failed
Docker image / build-and-push-image (push) Has been cancelled
Some checks failed
Docker image / build-and-push-image (push) Has been cancelled
get_public_key returns the signer pubkey without routing through pubkeyAllowed(), unlike every other NIP-46 method. #26 flagged this as an unaudited/ungated disclosure through the ACL seam and asked us to decide deliberately between gating it and documenting the exception. Verified against the live clients: gating it would BREAK production. lnbits' _ensure_policy (remote_bunker.py DEFAULT_POLICY_RULES + DEFAULT_POLICY_METHODS_NO_KIND) grants only sign_event(kinds) + the four nip04/44 crypto methods — no get_public_key rule — and the client calls get_public_key as a spec-mandated, hardcoded post-connect session step (nip46_bunker_client.py connect()). Routing it through checkIfPubkeyAllowed would return `undefined`, dropping that call onto the admin-approval path and stalling session establishment → "signer unavailable" (the #41 outage class). So the correct resolution is #26's option A (accept + document): the pubkey isn't secret, NIP-46 mandates it ungated during session setup, and the clients carry no grant for it. Make the exception explicit and load-bearing in the code so a future refactor doesn't "helpfully" gate it and reintroduce the outage. No behavior change; tsc-clean; test:nip46 green. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
commit
d35b98a8c9
|
|
@ -104,6 +104,29 @@ export class Backend {
|
|||
return ok ? "pong" : undefined;
|
||||
}
|
||||
case "get_public_key":
|
||||
// Intentionally UNGATED — do NOT route through pubkeyAllowed().
|
||||
// This is a deliberate, documented exception to the "every
|
||||
// request hits the ACL seam" rule (aiolabs/nsecbunkerd#26), not
|
||||
// an oversight, and it is load-bearing — gating it breaks the
|
||||
// live clients:
|
||||
// 1. The pubkey isn't secret — it's the identity the bunker
|
||||
// openly signs as; disclosure leaks nothing a relay
|
||||
// observer couldn't already derive from published events.
|
||||
// 2. NIP-46 mandates get_public_key as part of session
|
||||
// establishment (46.md overview step 5: the client requests
|
||||
// get_public_key to learn user-pubkey), called immediately
|
||||
// post-connect, before any policy-bearing method.
|
||||
// 3. lnbits' `_ensure_policy` (remote_bunker.py
|
||||
// DEFAULT_POLICY_RULES + DEFAULT_POLICY_METHODS_NO_KIND)
|
||||
// grants only sign_event(kinds) + nip04/44 encrypt/decrypt
|
||||
// — there is NO get_public_key rule — so routing it through
|
||||
// checkIfPubkeyAllowed would return `undefined`, and the
|
||||
// client's spec-mandated post-connect get_public_key call
|
||||
// would stall on the admin-approval path → "signer
|
||||
// unavailable" (the outage class #41 addressed).
|
||||
// Per-app gating/audit of identity disclosure, if ever wanted,
|
||||
// must be co-designed with a policy rule the clients actually
|
||||
// plant. See #26 for the full prior-art survey (NDK vs rust-nostr).
|
||||
return this.pubkey;
|
||||
case "sign_event": {
|
||||
const [eventString] = params;
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue