MLabs - Grumplestiltskin 2: Pasta Curve Builtins and Kimchi PLookup Proof-of-Concept
MLabs LTD
ID: 69fcab8de1ab5d003a0cab66
Zero-knowledge proofs (ZKPs) let one party verify that another knows something — or has performed some computation — without revealing the underlying data. They're foundational to privacy-preserving applications across blockchains. CIP-381 already brought basic ZKP support to Cardano via the BLS12-381 elliptic curve builtins, but the proof systems built on it are limited by the curve's properties: proofs are large, verification is slow, and many of the most efficient modern proof systems aren't usable at all. Mina's Kimchi proof system shows what better looks like. It uses a different curve family — the Pasta curves (Pallas and Vesta) — designed for recursive composition, where one proof can verify another. Combined with a technique called PLookup, it produces dramatically more efficient proofs for many practical use cases, including table-based privacy applications. Cardano cannot use this on-chain today because the necessary curve operations aren't available as Plutus Core builtins. This proposal closes that gap: A CIP describing UPLC builtins for the Pallas and Vesta curves, modeled on the existing CIP-381 BLS12-381 interface. An implementation in cardano-base (using the well-established arkworks-rs library) and the corresponding builtins in Plutus Core. A proof-of-concept verifier for Kimchi-variant PLookup proofs, built using the YTxP framework MLabs delivered under Catalyst Fund 12, demonstrating the new capabilities end-to-end. This continues from the original Grumplestiltskin project (a 2025 Cardano Treasury proposal that delivered a parameterized curve framework in Plutarch). Where Grumplestiltskin 1 made any curve usable in Plutarch with reasonable performance, Grumplestiltskin 2 adds specific high-value curves at the protocol level — much faster, but only for the curves that matter most. The two efforts are complementary.
Scorecard
How this score works
Each criterion is worth between 1 and 4 points. The score is the points earned out of the points on criteria that have been answered — shown as a percentage. Green is 75% or higher, Amber is 50–74%, and Red is below 50%.
Some criteria are checked automatically against the proposal data (open source, doxxed team, treasury return clauses, etc.). Others are human-judgment calls — value for money, public good, whether the deliverables are realistic. Those stay blank until a DRep ticks them.
Criteria can also be marked Not applicablewhen they don't fit the proposal — for example, "open source" doesn't apply to a DAO governance proposal with no software output. Those are excluded from the score entirely, neither helping nor hurting.
Green requires 80% coverage. If less than 80% of the applicable criteria (by points) have been answered, the verdict stays at Amber — "pending review" — even when every answered criterion passes. This proposal's coverage is currently 68% (13 of 19 applicable criteria answered).
Vote intent
Export & audit
↓ Download scorecard.mdRaw payload · Treasury (false) / Admin (intersect)
{
"_meta": {
"schema_version": "1.2.0",
"ingested_at": "2026-05-11",
"source_url": "https://hydra.intersectmbo.org",
"proposal_id": "69fcab8de1ab5d003a0cab66",
"title": "MLabs - Grumplestiltskin 2: Pasta Curve Builtins and Kimchi PLookup Proof-of-Concept",
"proposer": "MLabs LTD",
"administrator": "Intersect",
"estimated_duration_months": 8,
"requested_budget_ada": 1144433,
"requested_budget_usd": 205998,
"ada_usd_rate_proposer": 0.18
},
"identity": {
"doxxed": true,
"doxxed_notes": "MLabs LTD, IOG Plutus Partner. Public technical team across CTL, Plutarch, prior Plutus Core primitive contributions (secp256k1, bit-level manipulation), and Grumplestiltskin 1.",
"doxxed_set_by": "ai"
},
"metadata": {
"open_source": "yes",
"open_source_notes": "All deliverables stated as MIT-licensed or equivalent. CIP, cardano-base PR, Plutus Core PR, YTxP-based validator all public open-source.",
"open_source_set_by": "ai",
"has_prior_delivery": true,
"has_prior_delivery_notes": "MLabs is the only org outside IOG to have shipped Plutus Core primitives to mainnet (secp256k1/ECDSA and bit-level manipulation primitives). Grumplestiltskin 1 (2025 Cardano Treasury Withdrawal Proposal — parameterized elliptic curve framework in Plutarch) — M1 and M2 delivered, M3 implementation-complete pending third-party assurance sign-off due to year-end auditor delays. 44 funded Catalyst proposals overall. Plutarch and CTL continuously maintained.",
"has_prior_delivery_set_by": "ai",
"admin_model": "intersect",
"admin_model_set_by": "ai",
"duplicate_of_existing_solution": {
"value": "partial",
"confidence": "medium",
"set_by": "ai",
"notes": "Complementary continuation of Grumplestiltskin 1 (parameterized any-curve Plutarch framework). G2 adds specific Pallas/Vesta curves at the protocol level for higher performance. The technical critique from commenter 'perturbing' on the proposal raises a substantive question: the Midnight team has already forked Halo2 to work over BLS12-381 (the existing CIP-381 curve) and a 2025 paper from Midnight demonstrates efficient foreign field arithmetic in PLONK that enables recursion on BLS12-381 without needing the Pasta cycle. If that work proves out, much of the recursion-related value proposition of Pasta curves on Cardano is delivered via Midnight's existing BLS12-381 work — making the Pasta curve builtins less essential than the proposal claims. The proposal does not address this alternative path."
}
},
"ask": {
"usd_peg_vs_market": "conservative",
"usd_peg_vs_market_notes": "Uses 0.18 ADA/USD, well below the ~0.25 market baseline. Consistent with MLabs's batch peg.",
"usd_peg_vs_market_set_by": "ai",
"milestone_payments_disclosed": true,
"milestone_payments_disclosed_set_by": "ai",
"milestone_schedule_disclosed": "amounts_and_dates",
"milestone_schedule_disclosed_notes": "Five sequential milestones with week counts: M1 (2 weeks CIP), M2 (4 weeks cardano-base), M3 (2 weeks Plutus Core), M4 (2 weeks Kimchi prep), M5 (3 weeks validator implementation). Single WP with one line item for engineering hours. Eryx cited as cryptography-specialist auditor for milestone independent audit before payment release.",
"milestone_schedule_disclosed_set_by": "ai"
},
"treasury_return": {
"has_return_clause": false,
"has_return_clause_set_by": "ai",
"treasury_favourability": "weak",
"treasury_favourability_notes": "Standard MLabs delivery model — milestone-based, single labor line item, no revenue share or unused-funds return. Open-source delivery is the public-good return mechanism.",
"treasury_favourability_set_by": "ai"
},
"relationships": {
"competes_with": [],
"competes_with_notes": "No direct in-batch competitor. Note however the substantive technical critique from commenter 'perturbing' suggesting that Midnight's existing BLS12-381 foreign-field-arithmetic work (already implemented and audited) addresses recursion needs without requiring Pasta curve builtins. This is a directional disagreement about the technical path rather than a competing proposal.",
"competes_with_set_by": "ai"
},
"ecosystem_demand": {
"evidence_level": "weak",
"evidence_level_confidence": "medium",
"evidence_level_set_by": "ai",
"active_users_known": false,
"active_users_examples": [],
"developer_demand_signals": [
"'Core ZK capabilities' explicitly named as a Pillar 1 focus area in Cardano Vision 2030",
"ZK technology is an active area of cryptographic research with educational and developer interest",
"Mina ecosystem demonstrates production usage of Pasta curves and Kimchi PLookup"
],
"adoption_risk_notes": "No named Cardano-side teams currently waiting on Pasta curve builtins to ship a ZK application. Demand is structural ('Cardano lacks these primitives, so anyone building ZK on Cardano is constrained') rather than expressed (named teams have committed to using these once available). The deliverables are deliberately framed as foundational primitives + proof-of-concept rather than production-ready system — appropriate for foundational work but means adoption is multi-quarter post-delivery at best. The technical critique from 'perturbing' specifically argues that the alternative path (Midnight BLS12-381 + foreign-field arithmetic) is already implemented and audited, which if correct would mean Cardano can do the same recursive ZK work without this proposal. Rated weak because: (1) no named adopters; (2) credible alternative technical path exists; (3) PoC scope means even successful delivery does not produce immediate ecosystem usage."
},
"incumbents": {
"has_existing_solution": "partial",
"has_existing_solution_confidence": "medium",
"has_existing_solution_set_by": "ai",
"existing_solutions": [
{
"name": "CIP-381 BLS12-381 builtins",
"status": "shipped",
"overlap": "partial",
"notes": "Existing curve infrastructure on Cardano. Proposal positions Pasta as complementary rather than replacing, but for the recursion use case the relationship is more competitive — see Midnight critique below."
},
{
"name": "Grumplestiltskin 1 (MLabs)",
"status": "shipped",
"overlap": "partial",
"notes": "Parameterized any-curve framework in Plutarch — proposers' own prior work. G2 is presented as complementary (specific-curves-at-protocol-level vs any-curve-in-Plutarch) rather than duplicative."
},
{
"name": "Midnight Halo2 fork over BLS12-381 with foreign field arithmetic",
"url": "https://github.com/midnightntwrk",
"status": "shipped",
"overlap": "substantial",
"notes": "Per the on-proposal technical critique from commenter 'perturbing': Midnight has forked PSE Halo2 to work over BLS12-381 (Cardano's existing curve), and a 2025 Midnight paper demonstrates efficient foreign field arithmetic in PLONK that enables recursion on BLS12-381 without requiring the Pasta cycle. This work is already implemented and audited per the commenter. If correct, this provides an alternative path to recursive ZK on Cardano that does not require Pasta curve builtins. The proposal does not engage with this alternative."
}
],
"differentiation_claim": "Adds Pasta (Pallas/Vesta) curve operations as Plutus Core builtins for protocol-level performance. Enables Kimchi PLookup proofs and Mina-style recursive composition. Complementary to Grumplestiltskin 1's parameterized framework — fast specific curves vs slower any curve.",
"differentiation_credibility": "medium",
"differentiation_credibility_notes": "MLabs has unambiguous technical credibility for the delivery: only non-IOG team to ship Plutus Core primitives to mainnet, owner of Plutarch and YTxP. The delivery pattern (CIP, cardano-base PR, Plutus Core PR, PoC validator) is well-rehearsed. However, the substantive technical critique on the proposal page from 'perturbing' is exactly the kind of cryptography-domain expert input that reviewers should weigh carefully — it argues the alternative BLS12-381 path is already further along. Differentiation rated medium because the engineering credibility is high but the technical strategy question (Pasta vs BLS12-381+FFA) is not addressed in the proposal."
},
"risk_flags": {
"aggressive_peg": false,
"self_administered_no_oversight": false,
"incumbent_already_delivering": false,
"no_evidence_of_demand": true,
"competes_with_in_batch": false,
"concentration_risk": false
},
"id": "hydra-mlabs-grumplestiltskin-2-pasta-curves",
"source": {
"hydra_id": "69fcab8de1ab5d003a0cab66",
"channels": [
"hydra"
]
}
}