Kaspa ($KAS) core R&D - what developers discussed, debated, and shipped. May 28-June 1, 2026.
TN10 Hardening Activates Cleanly
The final TN10 ZK-hardening activation appears to have gone through without drama. Michael Sutton posted "clean activation" after the May 28 activation point, then immediately moved the discussion toward merging the finalized specs and follow-up cleanup. (source) (source)
That matters because the TN10 cycle was not just a testnet ceremony. It was the last short-loop validation pass for the Groth16/Risc0 verifier hardening, public-input pricing, SMT metadata, and KIP-21 inactivity shortcut before those changes were folded back into the main Toccata branch. (source) (PR #1026)
The KIP Specs Catch Up to the Code
Sutton also pushed a major KIP-21 update to align the document with the finalized implementation and make it more readable as a consensus spec. The update split proof/accessor details into companion docs, corrected the activity digest and tree formulas, added pruning-point bootstrap requirements, documented post-KIP-21 subnetwork and lane-ID rules, and added the inactivity shortcut definition. (source) (KIP PR #36)
KIP-20 was fixed and finalized the same day, and coderofstuff later confirmed that KIP-20 and KIP-21 were merged. KIP-16 also moved through a small TN10-status activation fix before merge, leaving KIP-17 as the remaining pending spec in that immediate batch. (source) (source) (source) (source) (source) (KIP-20 PR #35) (KIP-21 PR #36)
The next spec work is script pricing. Sutton said he was working on a complementary script-pricing KIP, then pushed fixes and updates to KIP-17 based on coderofstuff's fixes and IzioDev's review comments. (source) (source) (source) (KIP-17 PR #32)
Toccata Absorbs the Hardening Work
On May 31, Sutton merged Rusty Kaspa PR #1026 and opened PR #1027, describing the next step as merging all hardening logic cleanly into toccata. PR #1026 removed the temporary TN10 hardening activation split and made the hardened Groth16/Risc0 and seqcommit activity-root paths unconditional under Toccata. (source) (PR #1026)
PR #1027 then landed on June 1. Sutton said it ports the finalized TN10-tested Toccata ZK hardening changes back into toccata, without the TN10-only activation scaffolding, because mainnet Toccata has not activated yet. The PR finalizes stricter Groth16 validation, Risc0 witness-size bounds, per-public-input pricing, the KIP-21 seqcommit inactivity shortcut, and the related SMT metadata plus pruning-point/IBD sync behavior. (source) (PR #1027)
Low-Fee Relay Gets an Interim Path
The fee-policy transition also moved from discussion to code. coderofstuff said the interim period made sense because not everyone upgrades at once, and because he wanted to avoid the support problem of users asking why a transaction was taking too long while developers were focused on the Toccata hard fork. (source)
IzioDev clarified the edge case: the new minimum-fee threshold is a mempool policy issue, not a rule that invalidates already mined transactions. coderofstuff agreed: "Mempool only, yes. Mined txs have no concept of \"min fee threshold\"." (source) (source)
That discussion became Rusty Kaspa PR #1022, which allows P2P-relayed low-fee transactions before Toccata activation while updated nodes can still reject low-fee direct RPC submissions. Sutton approved the PR on June 1, and coderofstuff merged it shortly after. (source) (source) (PR #1022)
vProgs Settlement Testing Starts Surfacing
Hans marked vProgs PR #59 ready for review. The PR fixes a state-handling issue discovered by Maxim, improves root-computation performance, and addresses additional soundness bugs found while working on the SMT proof system. (source) (vProgs PR #59)
Sutton then asked whether it was fair to report that the L1 side was being tested and revalidated through a real based-app covenant, including guest activity and inactivity proving paths. Maxim answered that activity paths work on simnet in end-to-end tests, but inactivity paths had probably not been tested yet. Hans said the team is testing the full settlement pipeline and will report merged PRs from now on, while using unit and integration tests to iterate through bugs and missing pieces. (source) (source) (source) (source)
Wallet Standards Get Their Own KIP Direction
Wallet work also got a standards framing. Demisrael summarized kaspa-cli as a BIP-44 HD wallet and the multisig feature as directly BIP-87-style multisig, with the caveat that the old Go wallet uses derivation purpose 45 while BIP-87 uses 87. He proposed drafting KIP-44 and KIP-45 to document Kaspa wallet standards as an "architectural absorption" of BIPs while preserving Kaspa-specific behavior. (source)
All sources link to public messages in the Kaspa Core R&D (public) Telegram channel.