FD
← All projectsFoundation

Passport firmware

Firmware for Foundation Passport Bitcoin signing devices.

BitcoinHardware walletsNormal
Repository coverage

263 commits in the local evidence base

Every captured commit receives deterministic security triage and a separate communication-quality score. Security candidates and broader second-pass signals receive full-patch Ollama analysis.

102security candidates79second-pass queue215AI analyses
42commits · 30 days
128commits · 60 days
194commits · 180 days
256commits · 365 days
Backfill bands
Aug 5 → Feb 612 seen0 candidatesComplete
Feb 6 → Jun 659 seen3 candidatesComplete
Jun 6 → Jul 64 seen0 candidatesComplete
Jul 6 → Aug 52 seen0 candidatesComplete
Commit communication

Does the history explain itself?

Message quality measures whether a commit identifies its scope, purpose, rationale, testing, and supporting references. It does not change the security-severity score.

61/100 average clarity
16Strong · 80–100
116Adequate · 60–79
118Thin · 40–59
13Opaque · 0–39
4security candidates with opaque commit messaging
Read the scoring rubric →
Developer activity

Who is changing the project?

Public Git author strings; identities are not independently verified.

DeveloperCommitsCandidatesAnalyzedHigh riskMessage avg.
Jacksper13321424466
Matt Gleason841868351
Jack603454261
mjg-foundation693251269
Jean-Pierre De Jesus DIAZ828061
Ken Carpenter313048
QnA111060
dependabot[bot]505076
BitcoinQnA101045
Analysis record

Published AI watches

Last scanned 43 minutes ago

Informational 12 AI analysisMessage 78 · Adequate
FD FoundationPassport firmware BitcoinHardware wallets

SFT-8696: remove the Security Check from onboarding

This commit removes the 'Security Check' step from the initial device setup (onboarding) for the Passport hardware wallet. To avoid breaking devices that were mid-setup during a firmware update, the old step is kept as a harmless pass-thro…

8d77ba38by Jack+153−124 files
No security note in commit
Low 37 AI analysisMessage 73 · Adequate
FD FoundationPassport firmware BitcoinHardware wallets

Merge pull request #753 from Foundation-Devices/sft-8680-fail-to-sign-dynamic-taproot-qr-message

This commit adds support for signing BIP-322 message requests that wallets like Sparrow send as a PSBT (a Bitcoin transaction-like data format). It is a feature addition, not a reported vulnerability fix. The code includes careful validati…

New input parser for externally supplied PSBT data with explicit structural validationRejects non-Taproot BIP-322 requests and unsafe sighash typesBinds the user-reviewed message and address to the transaction being signed
5e98b6abby mjg-foundation+722−3011 files
No security note in commit
Moderate 59 AI analysisMessage 50 · Thin
FD FoundationPassport firmware BitcoinHardware wallets

SFT-8680: check derivation path length before indexing into it

This commit fixes a bug in the Passport hardware wallet's BIP-322 message signing code. Previously, when parsing a PSBT (a Bitcoin data format), the code could read past the end of an empty or too-short 'derivation path' field because it c…

Out-of-bounds read / index error on untrusted PSBT inputValidation order bug: first-byte check before length checkBIP-322 signing request parsing hardened
cec3925bby Matt Gleason+8−43 files
No security note in commit
Informational 22 AI analysisMessage 73 · Adequate
FD FoundationPassport firmware BitcoinHardware wallets

Merge pull request #754 from Foundation-Devices/sft-8681-passport-core-sign-taproot-messages-with

This commit changes how the Passport hardware wallet signs Bitcoin Taproot text messages. Previously it used SIGHASH_DEFAULT (the Taproot default), but now it uses SIGHASH_ALL. The change is for compatibility: the commit message says Sparr…

Change in cryptographic signature hash type (SIGHASH_DEFAULT -> SIGHASH_ALL)Comment explicitly references verifier compatibility and Sparrow wallet acceptanceUnit test asserts new signature encoding and verifies against an external test vector
87f43de1by mjg-foundation+22−172 files
No security note in commit
Low 41 AI analysisMessage 73 · Adequate
FD FoundationPassport firmware BitcoinHardware wallets

Merge pull request #752 from Foundation-Devices/sft-8674-passport-core-signing-flow-does-not-show-a-non

This commit adds on-screen warnings to the Passport hardware wallet when a user is reviewing a Bitcoin transaction that contains timelocks. Timelocks can prevent a transaction from confirming until a future date or until certain inputs hav…

UI warning for absolute time-based nLockTimeUI warning for BIP68 relative timelocks on version-2 transactionsHeight-based nLockTime intentionally not flagged to avoid false positives
5defe2c5by mjg-foundation+176−16 files
No security note in commit
Moderate 52 AI analysisMessage 85 · Strong
FD FoundationPassport firmware BitcoinHardware wallets

SFT-8674: warned about timelocks when reviewing a transaction

This commit adds on-screen warnings to the Passport hardware wallet when a Bitcoin transaction being reviewed includes a timelock that could delay confirmation. Before this change, a user could unknowingly sign a transaction that is blocke…

UI now warns users about transaction timelocks that delay confirmationPreviously undisclosed to the signer: absolute date locks and BIP68 relative locksHeight-based nLockTime intentionally not flagged to avoid warning fatigue
7cba5349by Jack+176−16 files
Vendor flagged security relevance
Moderate 59 AI analysisMessage 73 · Adequate
FD FoundationPassport firmware BitcoinHardware wallets

Merge pull request #747 from Foundation-Devices/sft-8653-passport-core-recheck-the-firmware-image-while

This update tightens the firmware update process on the Passport hardware wallet. Previously, the device verified the firmware file once, then copied it to internal flash memory without re-checking. A malicious or tampered file swapped in …

Time-of-check/time-of-use (TOCTOU) hardening in firmware update pathNew digest parameter threaded from verification task to staging taskStaging task now re-hashes firmware image and compares against expected digest
3449381fby mjg-foundation+68−264 files
No security note in commit
Moderate 60 AI analysisMessage 58 · Thin
FD FoundationPassport firmware BitcoinHardware wallets

Merge pull request #736 from Foundation-Devices/sft-8554-bootloader-setup-checks

This commit hardens how Passport's bootloader handles a sensitive factory key during initial device setup. Previously, the code trusted that erasing the key from SPI flash succeeded without checking, and it left a temporary copy of the key…

Sensitive key material (supply-chain validation key) is now explicitly zeroed from stack memory after useSPI flash erase of the SCV key is now verified by read-back instead of assumed successfulBootloader setup now aborts if the SCV key cannot be confirmed erased from SPI flash
ed62d448by mjg-foundation+245−55 files
No security note in commit
Moderate 62 AI analysisMessage 73 · Adequate
FD FoundationPassport firmware BitcoinHardware wallets

Merge pull request #735 from Foundation-Devices/sft-8555-update-restore-checks

This commit adds security checks when restoring a backup file to a Passport hardware wallet. Previously, multisig wallet configurations stored in a backup were restored without verifying that they actually belong to the seed being restored…

Validation of multisig backup records against restored seedDerivation verification for xpubs claiming backup root fingerprintDropping malformed records instead of failing entire restore
8dbae7e1by mjg-foundation+242−75 files
Vendor flagged security relevance
Moderate 60 AI analysisMessage 73 · Adequate
FD FoundationPassport firmware BitcoinHardware wallets

Merge pull request #748 from Foundation-Devices/sft-8654-passport-core-bounded-text-building-in-the

This commit replaces unsafe string-copying functions in the Passport hardware wallet's bootloader with a bounded helper that prevents text fields from overflowing their buffers. The change is defensive: it stops malicious or malformed firm…

Replacement of strcpy/strcat with bounded append_text helperHandling of non-NUL-terminated firmware header text fieldsBuffer overflow mitigation in bootloader UI message construction
e2d735f0by mjg-foundation+89−222 files
No security note in commit
Informational 15 AI analysisMessage 73 · Adequate
FD FoundationPassport firmware BitcoinHardware wallets

Merge pull request #732 from Foundation-Devices/sft-8545-native-taproot-key-signing

This commit adds a new, dedicated way for the Passport hardware wallet to sign Bitcoin Taproot transactions. It moves the cryptographic 'tweak' calculation from Python code into a Rust library and adds unit tests. There is nothing in the c…

5e3ed701by mjg-foundation+234−185 files
No security note in commit
Moderate 60 AI analysisMessage 73 · Adequate
FD FoundationPassport firmware BitcoinHardware wallets

Merge pull request #742 from Foundation-Devices/sft-8530-distinct-cosigners

This update to the Passport hardware wallet firmware adds a safety check when importing multi-signature wallet configurations. Previously, the same cosigner public key could be listed more than once under different fingerprints or address …

Adds duplicate-key detection during multisig wallet importCompares raw public-key x-coordinates rather than fingerprints or encoded xpub stringsPrevents a single key from being counted as multiple independent cosigners
92c68c63by mjg-foundation+69−03 files
Vendor flagged security relevance
Low 48 AI analysisMessage 73 · Adequate
FD FoundationPassport firmware BitcoinHardware wallets

Merge pull request #739 from Foundation-Devices/sft-8559-psbt-file-errors

This update fixes a bug where the Passport hardware wallet could freeze (hang) when reading a malformed or unreadable PSBT transaction file from a microSD card. Now the device reports a clear error instead of getting stuck, so users know s…

Denial-of-service / UI hang on malformed input now returns an errorUnhandled exception path now caught and reported instead of leaving task unresolvedMissing microSD card case explicitly reported
6a461136by mjg-foundation+144−62 files
No security note in commit
Low 34 AI analysisMessage 73 · Adequate
FD FoundationPassport firmware BitcoinHardware wallets

Merge pull request #740 from Foundation-Devices/sft-4482-update-card-missing

This commit fixes a user-experience bug in the Passport hardware wallet's firmware update flow. Previously, if the user removed the microSD card while the device was reading a firmware update file, the process would crash with an unhandled…

Adds exception handling for a removable-media failure during a privileged operation (firmware update)Prevents unhandled CardMissingError from aborting the update flow abruptlyDoes not alter firmware signature, timestamp, or header validation logic
06ec8d6aby mjg-foundation+109−23 files
No security note in commit
Moderate 59 AI analysisMessage 73 · Adequate
FD FoundationPassport firmware BitcoinHardware wallets

Merge pull request #741 from Foundation-Devices/sft-737-multisig-match-n

This firmware update fixes a bug in how the Passport hardware wallet matches multisignature (multisig) wallets. Previously, when looking for a saved wallet that matches the cosigners in a transaction, the device only checked that the cosig…

Incorrect multisig wallet matching could lead to signing with the wrong wallet policySubset-only cosigner matching allowed a smaller wallet to be conflated with a larger oneFix adds N filter to wallet candidate enumeration
d2d30a58by mjg-foundation+59−23 files
No security note in commit
Informational 15 AI analysisMessage 73 · Adequate
FD FoundationPassport firmware BitcoinHardware wallets

Merge pull request #746 from Foundation-Devices/sft-8260-evaluate-miniscript-no-compatible-path

This commit only changes the wording of an on-screen warning message shown to users when a Bitcoin transaction's timelock conditions are not yet met. It does not change any security logic, validation, or behavior. The old message said no s…

f6ab63ceby Jacksper13+5−42 files
No security note in commit
Informational 15 AI analysisMessage 73 · Adequate
FD FoundationPassport firmware BitcoinHardware wallets

SFT-8260: Update the policy no-compatible-path warning copy

This commit only changes the wording of an on-screen warning message shown to users when a Bitcoin transaction cannot be signed because no spending path's time-lock conditions are currently satisfied. It does not change any security logic,…

e024471eby Jack+5−42 files
No security note in commit
Informational 21 AI analysisMessage 58 · Thin
FD FoundationPassport firmware BitcoinHardware wallets

Merge pull request #731 from Foundation-Devices/SFT-974-miniscript-ui-followup

This commit is a follow-up to the miniscript wallet-policy user interface. It mostly fixes UI wording, memory cleanup, and policy display logic. The only potentially security-relevant change is that wallet policy IDs are now computed with …

Incremental SHA-256 digest used for policy ID computationRemoval of explicit `validate(miniscript, context)` call before compilationNew default-name derivation from descriptor checksum
83d1d576by mjg-foundation+415−5312 files
No security note in commit
Low 32 AI analysisMessage 50 · Thin
FD FoundationPassport firmware BitcoinHardware wallets

SFT-974: unwrapped miniscript parsing recursion, added default policy

This firmware update changes how the Passport hardware wallet handles complex Bitcoin wallet policies. The main visible change is replacing a recursive parser walk with an iterative one, which prevents the device from crashing due to exces…

Recursion in miniscript parsing replaced with explicit stack to avoid RecursionError on deeply nested templatesPolicy validation path changed: explicit `validate()` call removed, safety now enforced via `compile_miniscript()` during registrationDefault empty policy name now auto-populated from descriptor checksum to avoid unnamed policies
96b2aa9fby Matt Gleason+160−216 files
No security note in commit
Informational 20 AI analysisMessage 50 · Thin
FD FoundationPassport firmware BitcoinHardware wallets

SFT-974: improved miniscript UI, working on OOM issues

This commit is a routine feature and stability patch for the Passport hardware wallet's miniscript wallet-policy support. It improves on-screen descriptions of wallet policies, adds explicit memory cleanup (garbage collection) during polic…

Memory pressure mitigation via explicit garbage collection during policy operationsNew verify-address flow for registered wallet policiesUI clarity improvements for timelock and signer-role descriptions
cb7478a5by Matt Gleason+255−3210 files
No security note in commit
Repository ledger

Explore captured commits

Expand any commit for its author, full message, clarity score, changed files, triage signals, analysis, and source link.

AI review queuedSFT-8316: prevented null truncation in op_return messagesby Matt Gleason · 7bd78ead · Oct 23, 2026 · 2 filesMessage 50 · ThinLow 49Details
Commit message · Matt Gleason

SFT-8316: prevented null truncation in op_return messages

50/100 · ThinMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component! No meaningful explanatory body
Why it was queued
signing or wallet pathsecond-pass: security-sensitive path
AI analysis · Low 49/100

This commit fixes a display bug in Passport's Bitcoin signing flow. When a transaction contains an OP_RETURN message with a hidden NUL (zero-value) byte, the device's screen would silently cut off everything after that byte. An attacker could craft a message that looks short or harmless on the device, while the real message the user is approving is actually longer or different. The fix now detects NUL bytes and shows the message as raw hex instead, so the full content is visible.

AI review queuedMerge pull request #726 from Foundation-Devices/sft-7086-add-screen-confirming-amount-being-sent-to-op_returnby mjg-foundation · 8be0de19 · Oct 23, 2026 · 5 filesMessage 73 · AdequateModerate 60Details
Commit message · mjg-foundation

Merge pull request #726 from Foundation-Devices/sft-7086-add-screen-confirming-amount-being-sent-to-op_return

SFT-7086: label OP_RETURN outputs and warn when they carry value

73/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides an explanatory body✓ Links an issue, advisory, or supporting reference
Why it was queued
signing or wallet pathmerge-commit duplicate discountsecond-pass: security-sensitive path
AI analysis · Moderate 60/100

This commit adds safety warnings and clearer labels for a special kind of Bitcoin transaction output called OP_RETURN. Any bitcoins sent to an OP_RETURN output are permanently destroyed ('burned'). Previously, the wallet showed the amount and a user message but did not clearly tell the user the money was unspendable. The change also makes sure that if the OP_RETURN message contains unusual characters the device cannot display, it shows the raw bytes in hex instead of silently appearing blank, which could hide a hidden message or trick the user.

AI review queuedSFT-7086: fix hex import and test settings for the simulatorby Jack · 7938b404 · Oct 20, 2026 · 2 filesMessage 83 · StrongInformational 19Details
Commit message · Jack

SFT-7086: fix hex import and test settings for the simulator

ubinascii has no b2a_hex; take the hexlify alias utils already exports.
The unit harness leaves common.settings unset, so the warning test
supplies defaults for current_chain() and restores them afterwards.

83/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context✓ Mentions testing or verification
Why it was queued
signing or wallet pathsecond-pass: security-sensitive path
AI analysis · Informational 19/100

This commit fixes two small bugs that only affect the Passport hardware wallet's software simulator and its unit tests. It swaps one way of converting bytes to hex for another already-provided alias, and it gives the test harness default settings so a warning test can run. There is no indication this changes real-device behavior or fixes a security vulnerability.

Security candidateSFT-7086: label OP_RETURN outputs and warn when they carry valueby Jack · cc259d23 · Oct 20, 2026 · 5 filesMessage 73 · AdequateModerate 52Details
Commit message · Jack

SFT-7086: label OP_RETURN outputs and warn when they carry value

Transaction review showed an OP_RETURN output as an amount followed by a
message, which reads like an ordinary payment with a note, and nothing
named the output or said the amount can never be spent.

- Review shows "Destination: OP_RETURN" above the message.
- The message is never blank. No payload shows "(empty)", and a payload
with nothing the review font can draw (it only carries printable ASCII)
falls back to its bytes in hex.
- When OP_RETURN outputs carry value, a "Burned Amount" warning reports
the total. Review pages through each warning before signing. This warns
rather than blocks: a valued OP_RETURN is standard and is sometimes
used on purpose.

73/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context
Why it was queued
signing boundarysigning or wallet path
AI analysis · Moderate 52/100

This commit improves the user interface shown when reviewing Bitcoin transactions that include OP_RETURN outputs. OP_RETURN outputs permanently destroy any bitcoins sent to them. Previously, the wallet displayed these outputs like a normal payment with a note, which could mislead a user into thinking the funds were spendable. The change now clearly labels the output as OP_RETURN, never leaves the message field blank, falls back to hex for unprintable characters, and shows a warning when any value is being burned. It is a defensive UX fix rather than a cryptographic vulnerability patch.

AI review queuedMerge pull request #725 from Foundation-Devices/SFT-7375-paginate-transaction-detailsby mjg-foundation · e73a1c1c · Oct 15, 2026 · 7 filesMessage 73 · AdequateLow 29Details
Commit message · mjg-foundation

Merge pull request #725 from Foundation-Devices/SFT-7375-paginate-transaction-details

SFT-7375: switched from scrolling to pagination for transaction reviews

73/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides an explanatory body✓ Links an issue, advisory, or supporting reference
Why it was queued
signing or wallet pathmerge-commit duplicate discountsecond-pass: security-sensitive path
AI analysis · Low 29/100

This commit changes how the Passport hardware wallet displays transaction details during signing. Instead of one long scrolling screen, it now breaks the review into separate pages and requires the user to scroll to the bottom of each page before pressing forward. This is a user-interface hardening change meant to make sure people actually see all transaction details and warnings before approving. It is not a fix for a remote exploit or a cryptographic bug, but it does reduce the chance that a user accidentally approves a bad transaction because important information was hidden off-screen.

AI review queuedSFT-7375: switched from scrolling to pagination for transaction reviewsby Matt Gleason · bb6bf4f7 · Oct 13, 2026 · 7 filesMessage 50 · ThinLow 34Details
Commit message · Matt Gleason

SFT-7375: switched from scrolling to pagination for transaction reviews

50/100 · ThinMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component! No meaningful explanatory body
Why it was queued
signing or wallet pathsecond-pass: security-sensitive path
AI analysis · Low 34/100

This commit changes how the Passport hardware wallet displays transaction details during signing. Instead of showing all outputs, change, fees, and warnings in one long scrolling screen, it breaks them into separate paginated pages and requires the user to scroll each page to the bottom before pressing forward. This is a defensive UX improvement meant to ensure users actually see every output and warning before approving a Bitcoin transaction.

AI review queuedMerge pull request #677 from Foundation-Devices/SFT-974-miniscriptby mjg-foundation · 1a10714f · Oct 10, 2026 · 58 filesMessage 58 · ThinInformational 15Details
Commit message · mjg-foundation

Merge pull request #677 from Foundation-Devices/SFT-974-miniscript

SFT-974: miniscript

58/100 · ThinMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Links an issue, advisory, or supporting reference! No meaningful explanatory body
Why it was queued
signing or wallet pathmerge-commit duplicate discountsecond-pass: unusually broad changesecond-pass: security-sensitive path
AI analysis · Informational 15/100

This is a large feature commit that adds Miniscript wallet policy support to the Foundation Passport firmware. It introduces new modules for parsing and validating Miniscript policies, registering wallet policies via QR/microSD, deriving addresses, and signing PSBTs that use these policies. The changes also add UI flows, tests, and CI coverage. There is no indication in the commit message or diff that this is a security fix; it reads as a new feature implementation. No vulnerability is evident from the supplied materials.

Security candidateMerge pull request #724 from Foundation-Devices/sft-8255-passport-core-validate-unauthenticated-timestamp-readby mjg-foundation · d96f29d0 · Oct 9, 2026 · 5 filesMessage 58 · ThinModerate 69Details
Commit message · mjg-foundation

Merge pull request #724 from Foundation-Devices/sft-8255-passport-core-validate-unauthenticated-timestamp-read

Tighten firmware update state checks

58/100 · ThinMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Links an issue, advisory, or supporting reference! No meaningful explanatory body
Why it was queued
defensive validationboot or update pathmerge-commit duplicate discount
AI analysis · Moderate 69/100

This commit tightens how the Passport hardware wallet checks the firmware timestamp stored in its secure element during firmware updates. Previously, the device read the timestamp from encrypted storage but did not verify that the stored block was authentic and fresh. The patch adds a cryptographic check (pairing unlock + MAC verification) before trusting the timestamp, and makes production builds fail the update if the timestamp cannot be read or verified. It also adds tests that simulate tampered or stale secure-element responses. In plain terms: the wallet now makes sure the 'minimum allowed firmware version' number it uses during updates hasn't been forged or replayed by an attacker.

AI review queuedMerge pull request #713 from Foundation-Devices/SFT-7485-update-foundation-urby mjg-foundation · 61eb519e · Oct 9, 2026 · 3 filesMessage 58 · ThinLow 49Details
Commit message · mjg-foundation

Merge pull request #713 from Foundation-Devices/SFT-7485-update-foundation-ur

SFT-7485: update foundation-ur

58/100 · ThinMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Links an issue, advisory, or supporting reference! No meaningful explanatory body
Why it was queued
parser or protocol pathmerge-commit duplicate discountsecond-pass: security-sensitive path
AI analysis · Low 49/100

This commit updates a Rust dependency called foundation-ur from version 0.4 to 0.5 and adds a test showing that a malformed QR code fragment no longer corrupts the next valid scan. The change appears to fix a state-poisoning bug in the QR code decoder used by the Passport hardware wallet, but the actual fix is inside the external library, not visible in this diff.

Security candidateMerge pull request #714 from Foundation-Devices/SFT-7497-update-foundation-firmwareby mjg-foundation · 3ea15873 · Oct 9, 2026 · 4 filesMessage 58 · ThinLow 31Details
Commit message · mjg-foundation

Merge pull request #714 from Foundation-Devices/SFT-7497-update-foundation-firmware

SFT-7497: update foundation-firmware

58/100 · ThinMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Links an issue, advisory, or supporting reference! No meaningful explanatory body
Why it was queued
boot or update pathmerge-commit duplicate discount
AI analysis · Low 31/100

This commit updates the firmware verification library from version 0.2 to 0.3 and adjusts the surrounding Rust code to match. The visible changes are mostly defensive: adding safety documentation, marking C-callable functions as unsafe, and mapping a new error case so invalid firmware headers are reported more precisely. There is no direct evidence in the commit of a fixed security vulnerability, but the update to the underlying library could include undisclosed fixes. The commit itself does not explain why the library was updated.

Security candidateSFT-8255: report unavailable update state without a fatal errorby Jack · 4a6a8d69 · Oct 9, 2026 · 2 filesMessage 73 · AdequateLow 45Details
Commit message · Jack

SFT-8255: report unavailable update state without a fatal error

Handle the native state error on the header screen and stop the update without selecting a header. Cover the recovery path with the actual flow method.

73/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context
Why it was queued
boot or update path
AI analysis · Low 45/100

This commit changes the Passport hardware wallet's firmware update flow so that a specific internal error while reading the firmware header is shown to the user as a recoverable error message instead of crashing the device with a fatal error. It also adds a test to confirm the device can recover from this situation without getting stuck.

Security candidateSFT-8255: tighten firmware update state checksby Jack · b46ac00f · Oct 9, 2026 · 5 filesMessage 78 · AdequateModerate 59Details
Commit message · Jack

SFT-8255: tighten firmware update state checks

Validate stored firmware state before using it and share failure handling across update entry points. Preserve the existing data format and development-build behavior. Add focused native host coverage to CI.

78/100 · AdequateMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context✓ Mentions testing or verification
Why it was queued
defensive validationboot or update path
AI analysis · Moderate 59/100

This commit hardens how the Passport hardware wallet checks the stored firmware timestamp before allowing a firmware update. Previously, if the secure element returned a tampered or stale firmware timestamp, the device might have accepted it. Now it verifies the stored block with a fresh cryptographic check and treats a failed read as an error in production builds. The change also adds automated tests that simulate a malicious or misbehaving secure element.

AI review queuedMerge pull request #682 from Foundation-Devices/codex/sft-7352-blockstream-coreby mjg-foundation · 0059b02f · Oct 8, 2026 · 5 filesMessage 73 · AdequateInformational 15Details
Commit message · mjg-foundation

Merge pull request #682 from Foundation-Devices/codex/sft-7352-blockstream-core

SFT-7352: Add Blockstream crypto-account export

73/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides an explanatory body✓ Links an issue, advisory, or supporting reference
Why it was queued
signing or wallet pathmerge-commit duplicate discountsecond-pass: security-sensitive path
AI analysis · Informational 15/100

This commit adds a new feature to the Passport hardware wallet: exporting a Bitcoin account in a format compatible with Blockstream's Jade wallet. It is a pure feature addition with no visible bug fix, defensive hardening, or security patch. There is no indication in the commit that it addresses a vulnerability.

Security candidateMerge pull request #716 from Foundation-Devices/sft-7344-simulator-random-bytesby mjg-foundation · 3ddd757f · Oct 8, 2026 · 2 filesMessage 78 · AdequateLow 49Details
Commit message · mjg-foundation

Merge pull request #716 from Foundation-Devices/sft-7344-simulator-random-bytes

SFT-7344: read simulator entropy from the host

78/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides an explanatory body✓ Links an issue, advisory, or supporting reference✓ Names security-relevant behavior explicitly
Why it was queued
entropy or randomnessmerge-commit duplicate discount
AI analysis · Low 49/100

This commit fixes a weakness in the software simulator of the Passport hardware wallet. Previously, the simulator's random number generator was seeded with the system boot time, which could make simulated private keys or seeds guessable by someone who knew roughly when the simulator started. The change now reads randomness directly from the host computer's /dev/urandom. The commit message and documentation explicitly warn that the simulator is not a real hardware wallet and should never hold real funds.

AI review queuedSFT-7909: fixed lintby Matt Gleason · 74ed0598 · Oct 8, 2026 · 1 fileMessage 28 · OpaqueInformational 15Details
Commit message · Matt Gleason

SFT-7909: fixed lint

28/100 · OpaqueMessage clarity
✓ Subject identifies a change! No meaningful explanatory body
Why it was queued
second-pass: opaque commit message
AI analysis · Informational 15/100

This commit only reformats a single line of code in a test file to satisfy a code linter. It does not change what the code does, and it is not a security fix.

Security candidateMerge pull request #717 from Foundation-Devices/sft-8249-passport-core-psbt-error-text-loses-characters-to-lvglby mjg-foundation · f271fe38 · Oct 8, 2026 · 2 filesMessage 73 · AdequateLow 31Details
Commit message · mjg-foundation

Merge pull request #717 from Foundation-Devices/sft-8249-passport-core-psbt-error-text-loses-characters-to-lvgl

SFT-8249: escape PSBT error text before displaying it

73/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides an explanatory body✓ Links an issue, advisory, or supporting reference
Why it was queued
signing boundarysigning or wallet pathmerge-commit duplicate discount
AI analysis · Low 31/100

This commit fixes a display bug in the Passport hardware wallet where error messages about Bitcoin transaction files (PSBTs) could lose important text, such as input numbers, because the screen-rendering library treats the '#' character as a color-code marker. The fix escapes every '#' so the full error message is shown to the user. It is primarily a usability and information-integrity issue rather than a direct theft-of-funds vulnerability, because the error still blocks signing; however, a malformed PSBT could hide or alter the visible reason for refusal.

Security candidateSFT-7909: require a response long enough to fill the caller's bufferby Jack · d2fe7a8b · Oct 7, 2026 · 1 fileMessage 90 · StrongModerate 60Details
Commit message · Jack

SFT-7909: require a response long enough to fill the caller's buffer

Sizing the copy from len rather than from the transport count fixed the
overlong case and introduced a small one in the other direction. For OP_Info
with len=4, a response of exactly four bytes passes the minimum-length check
and skips the frame and CRC checks, so deserialize() writes tmp[0..3] and the
fixed-length copy reads tmp[1..4] -- one byte the transport never wrote --
then reports success.

Shown with a host harness compiling the real se_read(), se_read_response(),
deserialize(), check_crc() and se_crc16_chain() with only the transport and
logging replaced. A full response, then a truncated one, returned
data=11 22 33 ef with rc=0, where the transport had supplied only 11 22 33 and
ef was left in that stack slot by the preceding call.

Reject actual < len + 3 immediately before the copy, through the existing
retry path so a transient truncation gets another attempt. The guard sits
after the opcode checks on purpose: it only bites OP_Info, because for every
other opcode check_crc() requires tmp[0] == actual and the length check
requires tmp[0] == len + 3, which together already imply it. Keeping it there
also leaves the SE error-code path, which returns early on a four-byte
response, untouched.

Harness results with the guard: a full Info response still succeeds, the
truncated one returns -1 with the destination untouched, an overlong 263-byte
response still copies exactly len bytes without overflowing, and a normal
command's error response still surfaces last_error = 0xEE.

Overlong responses stay accepted rather than requiring actual == len + 3. The
comment in se_read_response() says framing is sometimes imperfect and junk can
arrive at the end, so an exact-equality gate risks refusing responses that work
today, and the overflow is already closed by copying len.

90/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context✓ Explains rationale or failure mode✓ Names security-relevant behavior explicitly
Why it was queued
memory safety
AI analysis · Moderate 60/100

This commit fixes a firmware bug in how the Passport hardware wallet reads replies from its secure chip. When a certain type of reply (an 'info' response) was shorter than expected, the code could copy unread bytes from leftover stack memory into the caller's buffer and report success. That could leak small amounts of whatever data happened to be on the stack, or return incorrect info about the secure chip. The fix adds a length check so short replies are rejected and retried instead of being silently accepted.

Security candidateSFT-7112: share one validation path between both message formatsby Jack · 027489af · Oct 7, 2026 · 2 filesMessage 83 · StrongModerate 57Details
Commit message · Jack

SFT-7112: share one validation path between both message formats

validate_lines() duplicated the text and path validation in each format
branch, and the empty-message guard was only ever added to the signmessage
branch. The legacy 2- and 3-line format therefore still assigned
self.text = self.lines[0] and called validate_sign_text() straight after, and
that opens with text[0] == ' ' under its default check_whitespace, so a file
beginning with a newline raised an uncaught IndexError.

Reachable from microSD and from QR: health_check_microsd_flow.py splits with
splitlines() and the guards above it only reject wholly empty content, not an
empty first line, and a 2-element list passes the length check.

Collapse the duplicated sequence into one validation path that runs for both
formats, with the empty check in front of it. Only the address-type derivation
stays branch-specific, since the legacy format carries its own type on the
third line while the Envoy format takes it from the path. That is the shape
originally proposed on the issue; splitting it per branch is what allowed one
copy to be fixed and the other missed.

Extend tests/unit/single_line_message.py with the legacy format: an empty first
line in both the 2- and 3-line shapes, in both signing modes, plus controls
that the legacy format still signs, still takes its address type from the third
line, and is still refused on an unusable path.

83/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context✓ Mentions testing or verification
Why it was queued
signing boundarydefensive validation
AI analysis · Moderate 57/100

This commit fixes a bug where a specially crafted message file could crash the Passport firmware with an IndexError. If a file loaded from microSD or scanned via QR began with a blank line, the firmware would try to read the first character of an empty message and crash. The fix consolidates the message validation so the same empty-message check applies to both the newer Envoy format and the older 2- or 3-line format. It also adds tests to prevent the bug from returning.

Security candidateSFT-8249: escape PSBT error text before displaying itby Jack · 34d0009b · Oct 7, 2026 · 2 filesMessage 81 · StrongLow 38Details
Commit message · Jack

SFT-8249: escape PSBT error text before displaying it

Labels render with LVGL recolor enabled, so a literal '#' in displayed text
opens a colour tag and the characters after it are consumed as a hex code
without being drawn. psbt.py raises a dozen messages shaped like "... for
input #3", and the index is the most useful part of them.

deef57bc escaped the transaction warnings for SFT-6782 and did not sweep the
error paths beside them, so 'Invalid PSBT: {}'.format(e) still lost characters.
Both sites now route through invalid_psbt_text(), which gives the escaping one
owner in this flow and makes it testable the way render_output() already is.

Scope is deliberately narrow. ErrorPage has 141 call sites, many passing
variables, so escaping inside ErrorPage would change behaviour across the whole
module tree at once. Error text arguably should never carry markup, but that is
a separate change and a separate review.

81/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context✓ Links an issue, advisory, or supporting reference
Why it was queued
signing boundarysigning or wallet path
AI analysis · Low 38/100

This commit fixes a display bug in the Passport hardware wallet where error messages about Bitcoin transaction files (PSBTs) could have important text hidden from the user. The device's screen library treats a '#' character as the start of a color command, so an error like 'problem with input #3' would show only 'problem with input ' and hide the number. The fix escapes the '#' so the full message appears. It is a user-interface reliability fix rather than a direct theft-of-funds vulnerability, but missing error details could mislead a user during transaction signing.

Security candidateSFT-7909: size the se_read copy from the caller's bufferby Jack · c5ec5134 · Oct 7, 2026 · 1 fileMessage 85 · StrongHigh 72Details
Commit message · Jack

SFT-7909: size the se_read copy from the caller's buffer

se_read() took its copy length from `actual`, the count of bytes that arrived on
the bus, rather than from `len`, the size of the buffer the caller supplied. tmp
only ever holds len + 3 bytes, because deserialize() clamps into it, so `actual`
describes neither the source nor the destination. Copy `len`.

check_crc() declared its length as uint8_t while its caller passes an int, so a
value above 255 wrapped before being compared against the response's count byte,
and a wrapped value can match a count that does not describe the response. Widen
the parameter. The count comparison still bounds it to a uint8_t before the CRC
walk, which is now said with a cast rather than left to the implicit conversion.

Verified by extracting check_crc() and se_crc16_chain() into a host harness and
running both signatures side by side: a well-formed seven-byte response passes
either way, a count of 263 is refused by the widened parameter and accepted by
the narrow one, and a mismatched count is refused by both.

85/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context✓ Explains rationale or failure mode
Why it was queued
signing boundary
AI analysis · High 72/100

This commit fixes two related bugs in the firmware that talks to the secure element chip inside the Passport hardware wallet. First, when reading data back from the secure element, the code was copying a number of bytes based on how many bytes arrived on the wire (`actual`) instead of how many bytes the caller's buffer could hold (`len`). That could overflow the caller's buffer if more bytes arrived than expected. Second, a CRC-check function used an 8-bit length variable, so any length above 255 would wrap around to a small number and could falsely match the response's count byte, letting an invalid or truncated response pass verification. The patch copies only `len` bytes and widens the length variable to an int, with an explicit cast only after it has been validated against the count byte.

AI review queuedSFT-7352: align Blockstream account export metadataby Matt Gleason · ebded2a2 · Oct 6, 2026 · 2 filesMessage 50 · ThinInformational 22Details
Commit message · Matt Gleason

SFT-7352: align Blockstream account export metadata

50/100 · ThinMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component! No meaningful explanatory body
Why it was queued
signing or wallet pathsecond-pass: security-sensitive path
AI analysis · Informational 22/100

This commit tweaks how the Passport hardware wallet exports a Bitcoin account to the Blockstream Jade wallet format. It changes the way the 'testnet/mainnet' coin type is encoded in the exported data so it matches what Jade's own software produces. There is no obvious security vulnerability here; it is a compatibility/alignment fix that could prevent the receiving wallet from misinterpreting the account type.

Security candidateSFT-7352: resolve test registration conflict and pin testnet CBORby QnA · f37f1219 · Oct 6, 2026 · 2 filesMessage 60 · AdequateInformational 12Details
Commit message · QnA

SFT-7352: resolve test registration conflict and pin testnet CBOR

60/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Mentions testing or verification! No meaningful explanatory body
Why it was queued
access control
AI analysis · Informational 12/100

This commit only reorganizes and strengthens a software test. It moves a test function to an earlier spot in a test list and replaces loose partial checks of test data with an exact full-value comparison. There is no change to the actual firmware code that runs on the Passport device, so users are not affected.

Security candidateMerge pull request #680 from Foundation-Devices/SFT-7355-harden-firmware-header-checkby mjg-foundation · 904a25f1 · Oct 6, 2026 · 10 filesMessage 58 · ThinModerate 59Details
Commit message · mjg-foundation

Merge pull request #680 from Foundation-Devices/SFT-7355-harden-firmware-header-check

SFT-7355: harden firmware header check

58/100 · ThinMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Links an issue, advisory, or supporting reference! No meaningful explanatory body
Why it was queued
defensive validationsigning or wallet pathboot or update pathmerge-commit duplicate discount
AI analysis · Moderate 59/100

This commit hardens how the Passport hardware wallet checks firmware update headers. It makes sure developer/user-signed firmware cannot carry a second public-key index or a second signature, and it prevents a 'time-of-check/time-of-use' race where the firmware file on the microSD card could be swapped after the user reviewed it but before it is copied to flash. The change is defensive: it closes a class of bugs where a partly trusted or attacker-modified header might be accepted, but there is no direct evidence in the commit that an actual exploit exists in the wild.

Security candidateMerge pull request #711 from Foundation-Devices/SFT-8212-ur-type-checkingby mjg-foundation · 6bacaeaf · Oct 6, 2026 · 8 filesMessage 58 · ThinLow 47Details
Commit message · mjg-foundation

Merge pull request #711 from Foundation-Devices/SFT-8212-ur-type-checking

SFT-8212: ur type checking

58/100 · ThinMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Links an issue, advisory, or supporting reference! No meaningful explanatory body
Why it was queued
cryptography-sensitive pathsigning or wallet pathmerge-commit duplicate discount
AI analysis · Low 47/100

This commit hardens how Passport firmware handles QR-code and UR (Uniform Resource) data when importing multisig wallet configurations. It replaces low-level 'mp_check_self' type assertions with explicit Python-style error checks, adds a new QR-scan flow that only accepts expected UR types, and adds unit tests to make sure wrong types are rejected. The changes reduce the chance that a malformed or unexpected QR/UR could crash the device or be misinterpreted as wallet data.

Security candidateMerge pull request #641 from Foundation-Devices/SFT-7112-single-line-signmessageby mjg-foundation · fb5bc630 · Oct 6, 2026 · 4 filesMessage 73 · AdequateLow 36Details
Commit message · mjg-foundation

Merge pull request #641 from Foundation-Devices/SFT-7112-single-line-signmessage

SFT-7112: handled signing the single-line message format

73/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides an explanatory body✓ Links an issue, advisory, or supporting reference
Why it was queued
signing boundarymerge-commit duplicate discount
AI analysis · Low 36/100

This commit adds support for a new single-line message-signing format used by Foundation's Envoy app. It also tightens parsing so that multi-line files split on all line endings (not just '\n') and so that embedded newlines in the single-line format are rejected rather than silently dropped. The change is a feature addition with defensive hardening, not a clear fix for an active vulnerability.