feat(debuglink): report the last confirmation text; reconcile the Solana LUT capability - #123
Open
BitHighlander wants to merge 10 commits into
Open
BitHighlander wants to merge 10 commits into
BitHighlander wants to merge 10 commits into
Conversation
dice_entropy alone is now the MIXED mode: the device commits its own 32-byte
draw as 24 BIP-39 words before the rolls are entered, then derives
seed = SHA256d("KK\x01SM" || draw || SHA256("KK\x01D" || rolls)). With
dice_only the derivation is seed = SHA256(rolls) and the draw is discarded,
matching Coldcard's Dice-Rolls-Only byte for byte.
The mode is a host-side selection so a wallet can explain what is coming --
99 rolls, and for MIXED 24 words to copy down -- before the ceremony starts.
The device still shows a consent screen naming the mode the host chose, so a
host cannot select dice-only silently. In both modes the host's EntropyAck is
consumed and its bytes dropped; the wire flow is otherwise unchanged.
dice_only without dice_entropy is rejected with a SyntaxError.
(cherry picked from commit 451e9a7)
Reports that ResetDevice.dice_only, the on-device consent screen and the tagged MIXED derivation are implemented. Needed because nanopb skips unknown fields: a host that sends dice_only to older firmware gets the older ceremony and a different wallet, with no error. Hosts and the test suite gate on this bit rather than on a version. (cherry picked from commit fbaf8ec)
…tocol python-keepkey's canonical branch already pins dd9c85d, which existed only on keepkey:reconcile/7.15-release. Merging it here makes one head serve the firmware and python-keepkey pins.
Tests could only read OLED pixels, so they asserted screen counts, which passed on unfixed firmware when raw glyphs paged like text. Reporting the formatted title and body lets them assert the exact text shown.
BitHighlander
added a commit
to BitHighlander/keepkey-firmware
that referenced
this pull request
Sep 25, 2026
…ot run Phase 0 of docs/security/HANDOFF-ERC7730-715-FORMATTERS.md (issue #861). The preload verifier accepted formatter kinds, display opcodes, path shapes and conditions that the signing runtime cannot execute, so a signed program could load and then fail after the user had approved the signer, the intent and earlier fields. Copilot found three narrower cases of this drift. - erc7730_capabilities.{h,c}: one table of what the runtime executes (intent at pc 0, raw fields without conditions, end; formatter kind 1 with one role-1 value; path source 1 with index steps; no conditions). The verifier and the runtime call the same predicates. - The verifier walks every value path against the ABI at preload: tuple index in range, fixed-array index in [-length, length), dynamic-array index in [-64, 64), and a leaf target. The walk table lives in the delegate-record buffer, unused until the bindings, so SRAM is unchanged (full reserve 18,336 B). - DebugLinkState.confirm_title/confirm_body (DEBUG_LINK only) record the last confirmation, so wire tests assert exact text instead of counts. - Pins device-protocol 5fec9e6 (keepkey/device-protocol#123 into up/release-protocol) and python-keepkey 67f561c (keepkey/python-keepkey#227 into reconcile/upstream-sync): both canonical heads plus these changes, and dp now also carries dd9c85d, which canonical pyk already pinned. Registry: 92 of 1,450 formats are signable (the plan said 94; two raw fields read the @.from container, which the runtime never captured). The device and the compiler's mirror agree on all 1,440 compiled formats.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Into the canonical 7.15 release protocol branch (PR #112).
1. Solana LUT capability reconciled. python-keepkey's canonical branch (keepkey/python-keepkey#197) and the 7.15 firmware stack already pin
dd9c85d(Features.supports_solana_lut_attestation = 29). That commit existed only onkeepkey:reconcile/7.15-releaseand was not on this branch. This merge makes one protocol head serve every pin.2.
DebugLinkState.confirm_title(16) andconfirm_body(17). These carry the title and body of the most recent confirmation, as the firmware formatted them and before pagination. They are filled only in DEBUG_LINK builds. Until now tests could read only OLED pixels, so they asserted screen counts, and a count-based test passed on unfixed firmware when raw glyphs paged the same as text. ERC-7730 formatter work (keepkey-firmware block 7, BitHighlander/keepkey-firmware#837, issue #861) needs exact-text assertions.Diff against
up/release-protocol:messages.proto, +9 lines. Field numbers 16 and 17 are unused on every branch.