aere-network/aere-docs
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
53f1b81117
aere-docs/AERE-STARK-DOC-CAVEAT-SWEEP.md
Aere Network dd1410c266 The unpublished line of work joins the sanitized public line 
The published line and the local line of this repository had no common
ancestor: the public one carried the hygiene pass (no host names, no internal
paths), the local one carried a month of corrections that never shipped. This
commit ports the local work onto the public line, keeping the public hygiene
wording wherever the two touched the same sentence, and keeping the public
version of AERE-CROSS-CLIENT-DETERMINISM.md entirely.

Carried: LICENSE/LICENSING corrections, VERIFY-POLICY.md,
CITATIONS-UNRESOLVED.md remeasured 2026-08-11, the 'audited' adjective removed
from next to Bouncy Castle, citation paths rewritten to published form, AIP-8,
the QA consolidation report, the second EIP validation pass, fork-height
corrections, the AereSink / threshold-factory correction, the forge test
floor, and the architecture-map updates.
2026-08-15 13:54:42 +03:00

12 KiB
Raw Blame History

STARK-verifier doc-accuracy sweep: Hypercube / retarget caveat

Date: 2026-07-19. Scope: docs only. No contract, precompile, or code touched. The source-of-truth finding lives in aerenew/docs/AERE-STARK-SP1-RECURSION-AIR-PORT-SPEC-SUMMARY.md (and the fuller AERE-STARK-SP1-RECURSION-AIR-PORT-SPEC.md), which were already correct and were NOT edited.

The correct framing applied everywhere

Aere has a real, conformance-confirmed native BabyBear + FRI STARK verifier skeleton (all 6 generic components validated byte-for-byte against the pinned Plonky3 0.4.3-succinct crates). It verifies BabyBear + FRI STARKs, for example Aere's own Plonky3 zk-circuits. It does NOT verify SP1 6.1.0 proofs: the pinned SP1 (facade = "=6.1.0", internals sp1-hypercube 6.3.1) is Hypercube (KoalaBear multilinear: BaseFold + Jagged/Stacked PCS + sumcheck-zerocheck + LogUp-GKR with a septic-curve digest), a different stack, so FRI and DEEP-ALI (the two biggest confirmed pieces) do not apply to it at all. Replacing the BN254 Groth16 wrap on real SP1 proofs is a separate ~22 to 32 person-week retarget (a founder decision). The real achievement is kept stated; only the scope is corrected.

Files changed (20) and the overclaim fixed in each

Explicit task files

  1. aerenew/docs/AERE-STARK-VERIFIER-PORT-SPEC.md

    • Overclaim: framed the whole ~11-16 person-week port as building a verifier for "an SP1 v6.1.0 INNER hash-based STARK proof: a Plonky3 ShardProof over the BabyBearPoseidon2 config" that "removes the BN254 Groth16 wrap ... from the SP1 trust chain"; called 0x0AE8 "the post-quantum replacement being built."
    • Fix: prominent SCOPE CAVEAT block at the very top; corrected Section 0 non-claims bullet; corrected the Section 1 target paragraph; added a scoped conclusion after Section 9 (the effort table is the cost of a BabyBear+FRI verifier, not an SP1 6.1.0 verifier).

  2. aerenew/docs/AERE-BUILD-LOOP-SUMMARY-2026-07-19.md

    • Overclaim risk: STARK row (item 2) could read as progress toward verifying SP1 proofs.
    • Fix: expanded the row with the scope caveat; explicitly states "does NOT let us verify SP1 proofs now," six generic components verify BabyBear+FRI (Aere's own circuits), retarget is separate.

  3. Per-component summaries (one-line scope caveat added at top of each; not rewritten wholesale):

    • AERE-STARK-BABYBEAR-FIELD-SUMMARY.md
    • AERE-STARK-POSEIDON2-SUMMARY.md
    • AERE-STARK-POSEIDON2-CONSTANTS-CONFIRMED.md
    • AERE-STARK-MMCS-SUMMARY.md
    • AERE-STARK-FRI-SUMMARY.md (notes SP1 6.1.0 uses BaseFold, not FRI)
    • AERE-STARK-TRANSCRIPT-SUMMARY.md
    • AERE-STARK-AIR-QUOTIENT-SUMMARY.md (notes the "six generic components confirmed" claim means a BabyBear+FRI skeleton, NOT an SP1 6.1.0 verifier; sumcheck-zerocheck not DEEP-ALI in SP1 6.1.0)
    • AERE-STARK-CORE-ADVANCE-SUMMARY.md
    • Overclaim in each: framed as a component of the "SP1/Plonky3 inner STARK-verify port."

  4. aere-research/pq-stark/README.md

    • Overclaim: title + opening claimed it "verifies SP1's INNER hash-based Plonky3/FRI STARK directly ... removes the one quantum-vulnerable link in the SP1 proof chain."
    • Fix: prominent SCOPE CAVEAT block at top; retitled to "BabyBear/Plonky3 FRI/STARK verify skeleton"; rescoped the opening paragraph to Aere's own Plonky3 circuits, not SP1 6.1.0.

Sweep: docs tree

  1. aerenew/docs/AERE-ENGINEERING-SECURITY-SPEC.md (Section 1.3 ZK-verification, quantum bullet)

    • Overclaim: "direct hash-based STARK verifier (verifying SP1's inner Plonky3 / FRI proof ...)".
    • Fix: replaced the parenthetical with the scope caveat (BabyBear+FRI targets Aere's own circuits; SP1 6.1.0 is Hypercube; removing the SP1 Groth16 wrap is a separate founder-gated retarget).
    • Note: A-9 ("BN254 is honest until replaced") and the guarantees table correctly state on-chain ZK verifiers are classical BN254 today; left TRUE and unchanged (points to the corrected 1.3).

  2. aerenew/docs/AERE-BENCHMARK-REPORT.md (Section B.3)

    • Overclaim: "a native precompile at 0x0AE8 that verifies SP1's INNER Plonky3 / FRI STARK directly."
    • Fix: rescoped to "a BabyBear / Plonky3 FRI STARK"; added a dedicated scope-caveat paragraph and adjusted the honest-status paragraph ("not a working verifier for SP1 proofs"; SP1 target is a different Hypercube system; retarget is separate).

  3. aerenew/docs/PQ-STARK-VERIFIER-PRECOMPILE-2026-07-18.md (design + spec, dated pre-finding)

    • Overclaim: whole premise, incl. title "direct SP1/Plonky3 inner FRI/STARK verification" and "The precompile verifies an SP1 v6.1.0 inner STARK proof - a Plonky3 ShardProof over the BabyBearPoseidon2 config."
    • Fix: prominent UPDATE / SCOPE CORRECTION banner at top stating the premise is incorrect for the pinned SP1 (Hypercube), directing readers to read "SP1 inner STARK / ShardProof" as "a BabyBear + FRI STARK, i.e. Aere's own Plonky3 circuits." Body left as the historical design record.

  4. aerenew/docs/PQ-STARK-VERIFIER-ACTIVATION-2026-07-18.md

    • Overclaim: P0 gate assumes activating a verifier of "a real SP1 v6.1.0 inner-STARK."
    • Fix: UPDATE / SCOPE CORRECTION banner at top (skeleton verifies Aere's own Plonky3 STARKs, not SP1 6.1.0; a founder-gated Hypercube retarget must precede any SP1-proof activation).

  5. aerenew/docs/AERE-AIP-PROCESS-AND-INDEX.md (AIP-16)

    • Overclaim: Decision said 0x0AE8 "verifies an SP1 inner-STARK (FRI over a Poseidon2 commitment)."
    • Fix: corrected the Decision to "verifies a hash-based STARK directly"; added a scope-caveat paragraph immediately below (BabyBear+FRI targets Aere's own circuits; SP1 6.1.0 is Hypercube; retarget is separate).

  6. aerenew/docs/AERE-EIP-COMPATIBILITY-MATRIX.md (0x0AE8 row)

    • Overclaim: name column "SP1 inner STARK verify."
    • Fix: relabeled "BabyBear/FRI STARK verify (skeleton)"; added the scope caveat to the notes cell.

  7. aerenew/docs/AERE-PROTOCOL-SPECIFICATION.md (Section 4.8.3, precompile table, intro list, recap)

    • Overclaim: 4.8.3 "direct on-chain verification of an SP1 (Plonky3) inner hash-based FRI/STARK proof ... remove the only quantum-vulnerable link (the BN254 Groth16 outer wrap) from the SP1 proof chain"; table row "SP1 inner STARK verify | Plonky3 FRI/STARK"; two label mentions (Section 4.6 list at line ~45 and the non-live recap at line ~670).
    • Fix: retitled 4.8.3 to "hash-based STARK verify," added the scope caveat to its body; relabeled the table row and both list mentions to "hash-based STARK verify (BabyBear/FRI, targets Aere's own Plonky3 circuits, NOT SP1 6.1.0 which is Hypercube; see 4.8.3)."

  8. aerenew/docs/AERE-PQ-FINALITY-CERTIFICATE.md

    • Overclaim: "SP1's core proof system is a FRI-based STARK" (factually wrong for SP1 6.1.0), and tying the AerePQStarkVerifier line to keeping SP1 proof verification quantum-safe.
    • Fix: corrected the clause (hash-based hence Shor-resistant reasoning kept; "FRI" label corrected: SP1 6.1.0 is Hypercube/BaseFold), and added that the 0x0AE8 BabyBear+FRI skeleton verifies Aere's OWN Plonky3 aggregation STARK, not an SP1 6.1.0 proof.

  9. aerenew/docs/AERE-XMSS-AGGREGATION-CIRCUIT.md

    • Overclaim: native-STARK-vs-Groth16 tradeoff implied the AerePQStarkVerifier line verifies the SP1-zkVM aggregation proof.
    • Fix: appended a scope caveat (AerePQStarkVerifier is BabyBear+FRI for Aere's own STARKs; the native-STARK option means an Aere-native BabyBear/FRI proof or a Hypercube retarget, not verifying an SP1 6.1.0 zkVM proof directly).

  10. aerenew/docs/AERE-QUANTUM-SAFE-NETWORKING.md (layer-3 quantum-layer map)

    • Overclaim: cell "PQ STARK verify, off BN254 | building" implied an imminent SP1-path replacement.
    • Fix: annotated the cell ("scope-corrected") and added a note under the table (skeleton is BabyBear+FRI for Aere's own circuits, does NOT yet verify SP1 6.1.0, retarget is separate, on-chain zk verifiers remain classical BN254 today).

  11. aerenew/docs/AERE-ARCHITECTURE-MAP.md (cross-cutting ZK-verification section)

    • Was already honest ("A post-quantum STARK verifier is roadmap"; on-chain verifiers classical BN254 today). Added a scope caveat for precision (0x0AE8 skeleton is BabyBear+FRI for Aere's own circuits; does NOT yet verify SP1 6.1.0; retarget separate; "not done or imminent"). The BN254-today statement was TRUE and left unchanged.

  12. aerenew/docs/AERE-QA-CONSOLIDATION-REPORT.md (precompile-consistency item 2)

    • The consistency snapshot propagated the "SP1 inner STARK verify" label.
    • Fix: relabeled 0x0AE8 to "hash-based STARK verify" and appended the scope caveat.

Sweep: pqc-fork tree

  1. aere-research/pq-stark/spec-mmcs-babybear.md

    • Overclaim: "the exact SP1 inner config" / "which MMCS Plonky3/SP1 use."
    • Fix: top scope caveat ("Plonky3/SP1" means Plonky3 0.4.3-succinct, Aere's own BabyBear+FRI stack, not SP1 6.1.0 Hypercube; confirmed for a BabyBear+FRI verifier, not an SP1 6.1.0 one).

  2. aere-research/pq-stark/spec-fri-babybear.md

    • Overclaim: "the exact FRI low-degree test Plonky3/SP1 use."
    • Fix: top scope caveat (SP1 6.1.0's inner recursion uses BaseFold not FRI; FRI does not apply to SP1 6.1.0; confirmed for Aere's own Plonky3 circuits).

  3. aere-research/pq-stark/spec-poseidon2-constants.md

    • Overclaim: constants framed as those "Plonky3/SP1 use."
    • Fix: top scope caveat (these are BabyBear constants; SP1 6.1.0 is KoalaBear/Hypercube and uses different constants; confirmed for a BabyBear+FRI verifier).

  4. aere-research/pq-stark/export-inner-stark-vector.md

    • Overclaim: whole premise "export the SP1 inner hash-based FRI/STARK proof (the shrink/compress Plonky3 shard proof over BabyBear + Poseidon2 + FRI) from the pinned SP1 v6.1.0 toolchain."
    • Fix: prominent SCOPE CORRECTION banner (SP1 6.1.0 exports a Hypercube/KoalaBear ShardProof, not a BabyBear+FRI ShardProof, which this skeleton does not verify; a real corpus for the current skeleton must come from Aere's own Plonky3 BabyBear+FRI provers; an SP1 6.1.0 corpus needs the Hypercube retarget first). Rescoped the opening paragraph accordingly.

Files reviewed, NO change needed (no overclaim found)

  • AERE-STARK-SP1-RECURSION-AIR-PORT-SPEC-SUMMARY.md, AERE-STARK-SP1-RECURSION-AIR-PORT-SPEC.md: source of truth; already document the Hypercube finding and the ~22-32 wk retarget correctly.
  • WHITEPAPER-V2.md, wp2-sections/02-architecture.md: their "verifies SP1 ... proofs" references are to the LIVE BN254 SP1 Groth16 gateway (0x9ca479...) via SP1VerifierGateway. Correct and TRUE; no native-STARK-verifies-SP1 claim. No change.
  • wp2-sections/04-zk.md, 06-mev-interop.md, 07-compliance.md: no 0x0AE8 / native-STARK-replaces- SP1 claim (only live-BN254 references). No change.
  • AERE-PQ-FINALITY-CERTIFICATE-SUMMARY.md (line ~53): "native hash-based STARK proof" is generic and honestly marked [VERIFY]/design-only; hash-based remains true for Hypercube; no FRI/SP1-verification overclaim. No change (parent certificate doc already caveated).
  • AERE-MIGRATION-TOOLKIT.md, AERE-NIST-VALIDATION-STATUS.md: only mention 0x0AE8 as testnet/non- live; no verification-capability claim. No change.
  • PQ-CONSENSUS-LIVE-READINESS-2026-07-18.md, ZK-PQ-FINALITY-2026-07-18.md, ZK-ANCHOR-FIXES-2026-07-17.md, PROVEN-RESULTS-2026-07-11.md, FRONTIER-PROGRAM-2026-07-12.md, VALIDITY-ROLLUP-SEQUENCER-PLAN.md, AUDIT-REPORT-2026-07-14.md: matched only the broad initial grep via live-BN254 SP1 gateway references; no native-STARK-verifies-SP1 / BN254-replacement claim. No change.

Guardrails honored

  • Docs only. No contract, precompile (0x0AE8 stays fail-closed), or other code touched.
  • Real achievement kept stated accurately (BabyBear+FRI skeleton is real and conformance-confirmed); only scope corrected. No overclaim in the other direction.
  • No em-dashes and no prose "--" introduced (verified; pre-existing em-dashes elsewhere left as-is).
  • Brand: "Aere Network" / "Aere" title case; token TICKER "AERE" uppercase; addresses/URLs untouched.

Confirmation

After the sweep, no doc implies the 0x0AE8 STARK verifier verifies SP1 6.1.0 proofs, replaces the BN254 Groth16 gateway for SP1, or is a "native PQ STARK verifier for SP1," without the Hypercube / retarget caveat. The on-chain ZK verifiers are still correctly stated as classical BN254 today.