Recognized players can't enter draws; alias sign-up unassemblable

Open
#320 0 comments 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
5/5
Estimated time
Over a week
Newbie friendliness
15/100
Issue type
Feature
Clarity
Mostly clear
Activity status
Stale
Tech stack
android, rust, typescript

Research direction

Start with pallets/game/src/lib.rs at the proof and alias-sign-up paths, then inspect pallets/airdrop/src/lib.rs and types.rs for the VRF context. Compare these with host-rust-core signing_host.rs and ring_vrf.rs, plus Android's MessageSigningContext. Done requires an accepted upstream chain- or host-side design; the issue states that product-sdk itself needs no code change.

Written by the indexing model from the issue text.

Description

Description

Game.sign_up_with_alias cannot be assembled by any product, and a recognized player cannot enter a game's prize draws by any route. Two independent blockers, neither fixable in this repo. Found while building #291 step 5; blocks #291 step 6. Part of #286.

Updated 2026-08-30. Both blockers re-verified, both still present. Three changes since filing: individuality#1300 merged without covering the airdrop context, the pallets moved to paritytech/individuality-community, and the escape hatch this issue asks not to ship is now proposed for merge. Paths below are individuality-community 1cbe91b.

Blocker 1: the airdrop VRF context

A recognized player must supply AirdropVrfs::Alias (pallets/game/src/lib.rs:2165, documented at types.rs:203). Its entries are ring-VRF proofs verified for blake2_256(AIRDROP_CONTEXT_BASE ++ event_id), where the base is the ASCII constant b"pop:polkadot.network/airdrop" (pallets/airdrop/src/lib.rs:93).

A host can only sign under a context it derives itself:

// host-rust-core, truapi-server/src/runtime/signing_host/ring_vrf.rs:215
input.extend_from_slice(b"product/");
input.extend_from_slice(context.product_id.as_bytes());
input.push(b'/');
input.extend_from_slice(&suffix);
blake2b_256(&input, None)

ProductProofContext is { product_id, suffix } and admits nothing else. DerivationIndex::Raw([u8; 32]) is not a way out — it is the suffix, inside the hashed preimage, not the output. The two preimages cannot be made to agree without a preimage attack.

The upstream fix landed and skipped this. individuality#1300 merged 2026-08-19 (part of #1247), moving six contexts onto the product-scoped derivation — including Index(4) people-airdrops. It did not touch pallets/airdrop. Note the trap: pallets/people-airdrops and pallets/airdrop are different pallets, and the one AirdropVrfs::Alias calls was not converted. The risk this issue named when opened is what happened. #1247 stays open for the remainder.

The open sub-question there is which contexts are enumerable enough for a u32 suffix. This one is not, as a whole — the event id is 32 bytes and varies per draw — though its pieces are: event_id = airdrop_event_id_base()(27) ++ airdrop_index(u8) ++ game_index(u32 BE).

Blocker 2: the statement-account proof

Unchanged since filing. Blocks the alias sign-up even with airdrops: None, so there is no partial path.

// pallets/game/src/lib.rs:1999
let msg = (Self::proof_of_ownership_msg_base(), &who)
    .using_encoded(sp_io::hashing::blake2_256);
ensure!(
    statement_account_signature.verify(&msg[..], &statement_account),
    Error::<T>::InvalidStatementAccountSignature,
);

The signature must be over the bare 32-byte hash, and the pallet unwraps nothing. The only raw-signing route an SDK has is PolkadotSigner.signBytes → the host's signRaw, and both hosts wrap unconditionally as <Bytes> ++ raw ++ </Bytes> — Rust core in signing_host.rs, Android via MessageSigningContext.generalUntrustedMessage(). A blake2 hash does not start and end with the markers, so the already-wrapped pass-through does not help.

Both hosts can sign unwrapped — Android's trustedContent() does, in four flows — but in every one the host builds the message; none takes bytes from a product. That is the whole point of the wrapper: product-supplied bytes signed bare could be a valid extrinsic. So the safe host-side shape is a first-class operation where the host builds the proof message from the alias it already holds, and the product supplies no bytes at all.

What works today

product-sdk#319 ships the account path and reports both blockers as AliasVrfsUnavailable and AccountVrfsNeedAnAccount, so a product finds out before submitting. That matters because a rejected sign-up costs a fee — Pays::No applies on success only, and the airdrop registration rides in the same extrinsic.

Works
Not recognized, sr25519, sign_up_with_account with draw entry yes, #319
Signed account, sign_up_with_account, no draw entry yes
Recognized player, draw entry, either sign-up call no, blocker 1
sign_up_with_alias, with or without draws no, blocker 2
Asks

Blocker 1 alone unblocks draw entry for a recognized player signing up with an account, which is the bigger win. Both are needed for the alias path.

  1. Preferred, chain. Derive the airdrop context the way score_context() and resources_context() were, so a product-scoped proof satisfies it. Matches the precedent #1300 set, and settles mob-rule too.
  2. Chain. Make the statement-account proof producible: accept the <Bytes>-wrapped preimage, or move the proof to something a host can produce.
  3. Or host. First-class operations for both, where the host builds the message and the context.

Neither blocker has an issue of its own upstream, which is the likeliest reason neither has moved. Both belong in individuality-community now. Who owns the chain-side change is the one answer that would get this moving.

Not a raw-context or raw-bytes escape hatch, please

Either would work mechanically. A raw context lets any product mint a proof in any chain context, defeating the isolation the product/ prefix enforces; unwrapped product bytes let a product collect a signature over what is really an extrinsic.

This is now a live proposal. host-rust-core#457 opened 2026-08-19 as a spike marked "not for merge"; as of 2026-08-27 it is out of draft, retitled spike( → feat(, grown from 41 lines to 8 files, and down from two gates to one — the environment variable is gone. Its body: "the hatch is on in every host until it is deleted." What remains is the raw: product id, which dotNS cannot issue. Its isolation design is good — two files, deletes in one go — but its stated motivation includes "one product (dim2.dot) proving into another product's context (people.dot)", which is precisely what the prefix prevents. If it merges in this shape, please keep a deletion issue open and keep raw: unissuable. Fixing the context in the chain removes the need for it.

Notes
  • The pallets moved repository. individuality-community is where they are developed; paritytech/individuality main has had nothing land since 2026-08-20. Both still carry a pallets/ tree, so a search in either looks authoritative.
  • Verified 2026-08-30 against individuality-community 1cbe91b, host-rust-core 4b38d782, polkadot-app-android-v2 c2a58d05b, product-sdk e74eb47.
  • Two products hit blocker 1. dim2-spa builds the same VRF transcript and only ever calls sign_up_with_account, so the two agree on everything except the arm neither can reach.
  • withAsPerson (#309) produces the person origin and names Game.sign_up_with_alias in its doc. It is right about the origin; the other arguments are what this issue is about.
  • individuality#1233 would widen sign_up_with_alias to lite-people aliases, triaged non-P0 — worth knowing that the call is unusable by products regardless of who may make it.
  • #291 step 7, the PeopleAirdrops path, is blocked separately: that context is not in the runtime's AccountContexts. Fixed 2026-08-24 by individuality-community#48.
  • Nothing here is fixable in product-sdk. What this repo gains afterwards is one more accepted context and one more producible signature, not new code.
Dominant language
TypeScript
Stars
14
Forks
5
Avg merge
1d 13h
Merged PRs (30d)
23

Getting set up

This project ships no dev container, Dockerfile or contributing guide, so setting up is up to you: start from its README, and see our first-contribution guide for the general steps.

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from paritytech/product-sdk

All issues in paritytech/product-sdk

Similar issues

More TypeScript issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.