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.
Message quality measures whether a commit identifies its scope, purpose, rationale, testing, and supporting references. It does not change the security-severity score.
This is a routine GitHub Actions workflow fix. It changes which Slack webhook secret is used for nightly test failure notifications and removes a now-unnecessary safety check. There is no security vulnerability in the change itself.
This commit only improves a test report for Trezor's clear-signing feature. It changes how skipped test cases are listed in a nightly summary, showing each skipped case's source file and calldata size instead of just counting them by reaso…
This commit only changes test scripts and documentation for the Trezor firmware clear-signing test suite. It makes the tests run more registry cases—specifically those where the developer definitions do not yet have a display format—so the…
This commit updates French and Portuguese translation strings in the Trezor firmware and refreshes the translation signature metadata. The changes are purely wording adjustments (for example, replacing abbreviations like 'nº' with the full…
This commit changes a code-coverage tracking map inside the Trezor firmware from a claimed capacity of 1,048,576 entries down to 65,536 entries. The commit message says the old capacity was never truly that large because the heapless libra…
No cryptographic or secret-handling code changedNo input parsing, deserialization, or memory allocator changesChange is in a debug/test instrumentation module (coverage)
This commit is a routine update of translated user-interface text strings for the Trezor hardware wallet. It adds new translations (for example, warnings about exporting public keys and Solana blind-signing prompts), updates existing wordi…
This commit only adds new test cases for Ethereum clear-signing array parsing. It does not change any production code, so it cannot introduce a new security vulnerability on its own. The tests verify that malformed arrays (out-of-bounds of…
test-only changeno production code modifiedadds defensive test coverage for ABI decoding edge cases
This commit hardens the Trezor firmware's task scheduler against fault-injection attacks. These attacks use physical techniques like voltage glitches or laser pulses to trick the device into executing code with higher privileges than inten…
Fault-injection hardening (FIH) explicitly mentioned in commit title and code commentsPrivilege-level transition split into multiple guarded stepsAdded runtime invariant check that only the kernel task may run privileged
This commit only removes an unused parameter called `is_dynamic` from test helper functions and from test data in three test files. It does not change any production firmware code, wallet behavior, or security logic. It is a cleanup of the…
This commit updates the data-format definitions used by Trezor devices for Ethereum 'clear signing' (showing human-readable details of smart-contract interactions). It adds support for more numeric types, new path-step structures, and extr…
Schema expansion increases attack surface for future clear-signing parser logic, but no new parser logic is present in this commit.Replacement of NotImplementedError with InvalidFormatDefinition is a defensive change that turns unimplemented paths into a handled error rather than a crash/DoS vector.New optional fields and deprecated old fields (is_dynamic, lookup_address) are backward-compatible changes.
This is a bug-fix patch in the Python Trezor client library. It changes how the software probes whether a connected device uses the older protocol v1. Previously, the probe tried to read and throw away leftover data, but that read could ac…
No security-relevant signal in commit message or diffFix is described as a connectivity/functional bug, not a vulnerabilityNo mention of CVE, researcher credit, or security advisory
This commit is a user-interface improvement, not a security fix. It changes the language-change confirmation prompt on Trezor devices so that it shows a human-readable language name (for example, 'French') instead of a raw code like 'fr-FR…
This commit updates Spanish and French translation strings in the Trezor firmware and refreshes the translation signature file. The changes are purely textual wording adjustments (for example, changing 'Mostrar palab.' to 'Mostrar palabras…
This commit hardens cryptographic code in Trezor firmware against two potential side-channel leaks. First, it simplifies how a secret condition value is checked inside an assert statement so the compiler produces only one branch instead of…
Side-channel hardening in constant-time cryptographic primitivesReplacement of double-evaluation volatile assert with single-branch assertElimination of secret-dependent memory access in elliptic-curve scalar multiplication
This commit adds a new nightly automated testing workflow for Ethereum 'clear signing' on Trezor hardware wallets. It does not change any wallet firmware code that users run; it only adds test scripts, CI configuration, documentation, and …
This commit is a routine code cleanup. It removes a small helper function called keccak_32 from a shared crypto utility file and places an equivalent helper directly inside the only test file that still uses it. There is no change to how t…
This commit removes support for a cryptocurrency called SmartCash from Trezor hardware wallets. It deletes coin definitions, support entries, and a custom cryptographic curve used only by SmartCash. There is no security vulnerability being…
No security-relevant signals detectedPure feature removal / code cleanup
This commit removes all support for the Groestlcoin cryptocurrency from Trezor firmware. It deletes coin definitions, the custom Groestl hash algorithm code, related address/transaction handling, and tests. There is no security vulnerabili…
No security-relevant change: pure feature removalNo new attack surface introducedNo vulnerability patched or created
This commit removes all support for the Tezos cryptocurrency from Trezor firmware, software libraries, and documentation. It is a feature-removal change, not a security fix or vulnerability patch. There is no indication in the commit that …
This update reduces the longest allowed screen-lock timeout on older Trezor devices from about 6 days down to 149 hours (just over 6 days). It makes the legacy firmware behave the same as newer Trezor firmware and prevents users from setti…
Bounds enforcement added for a configurable security-relevant timeoutAlignment of security policy between legacy and core firmware variantsTest coverage added for boundary values above and below new maximum
build(core): exclude meminfo-related code from FW builds
Saves ~7.3KB on T2T1.
Keep it for emulators.
[no changelog]
77/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Provides an explanatory body
Why it was queued
boot or update path
AI analysis · Informational 15/100
This commit is a routine firmware size optimization. It removes a developer-only memory debugging feature (meminfo) from production hardware builds to save about 7.3KB of flash storage, while keeping it available in emulator builds. There is no security vulnerability here.
Security candidatefeat(core): encapsulate select menu item in a structby Vojtěch Nevřela · f67410a6 · Sep 3, 2026 · 7 filesMessage 62 · AdequateInformational 15Details
Commit message · Vojtěch Nevřela
feat(core): encapsulate select menu item in a struct
[no changelog]
62/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope! No meaningful explanatory body
Why it was queued
cryptography-sensitive pathboot or update path
AI analysis · Informational 15/100
This commit is a straightforward internal code cleanup in the Trezor firmware's user-interface code. It replaces a raw two-value pair (text label + intent) with a named struct called SelectMenuItem. There is no change to user-visible behavior, no bug fix, and no security-related change.
✓ Descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope! No meaningful explanatory body
Why it was queued
boot or update path
AI analysis · Low 27/100
This commit adjusts memory layout alignment settings in the linker script used for Trezor's STM32F4 hardware wallet firmware. It moves a 512-byte alignment boundary from the end of one flash section to the end of another. The change is described by the vendor as fixing image size alignment on F4 devices. There is no explicit security claim in the commit, and the diff alone does not demonstrate a vulnerability or exploit path. It could plausibly relate to ensuring correct firmware image boundaries for secure boot or flashing, but that is speculative.
test: check `Features.max_passphrase_len` in device tests
Following 835efbb65caa8284c08da97e7378f5677eecb019, legacy and core have different passphrase limits.
[no changelog]
87/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Provides an explanatory body✓ Mentions testing or verification
Why it was queued
access controlauthentication path
AI analysis · Informational 15/100
This commit only adds a test assertion to verify that the maximum passphrase length reported by a Trezor device matches the expected value for its model family (50 characters for legacy devices, 128 for newer 'core' devices). It is a pure test-code change with no product behavior change and no security fix.
Security candidatechore: remove NEM and EOSby obrusvit · b6f408c7 · Sep 2, 2026 · 160 filesMessage 57 · ThinInformational 15Details
Commit message · obrusvit
chore: remove NEM and EOS
57/100 · ThinMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope! No meaningful explanatory body
Why it was queued
cryptography-sensitive pathsigning or wallet pathboot or update path
AI analysis · Informational 15/100
This commit removes all support for the NEM and EOS cryptocurrencies from Trezor firmware. It deletes the application code, protocol definitions, cryptographic helpers, translations, tests, and support entries for both coins. The change is a feature removal, not a security fix, and does not introduce any vulnerability.
Security candidatetest(core): use equality for passphrase click testsby Roman Zeyde · 2bd21719 · Sep 2, 2026 · 1 fileMessage 72 · AdequateInformational 15Details
Commit message · Roman Zeyde
test(core): use equality for passphrase click tests
[no changelog]
72/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Mentions testing or verification! No meaningful explanatory body
This commit only changes automated test code for the Trezor hardware wallet. It makes passphrase display checks stricter in a test file (using exact equality instead of substring matching). There is no change to the actual firmware, wallet logic, or any code that runs on a user's device. It cannot affect real users or introduce a security vulnerability.
refactor(core/eckhart): replace `String<99>` with `ShortString`
9695e1c2ef has increased `ShortString` size to 128.
[no changelog]
77/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Provides an explanatory body
Why it was queued
boot or update pathauthentication path
AI analysis · Informational 15/100
This is a small internal code cleanup in the Trezor firmware's PIN entry screen. It swaps one fixed-size text container (`String<99>`) for another (`ShortString`) that was already enlarged to 128 bytes in a prior commit. The behavior and safety limits remain the same, and there is no indication of a security fix.
Security candidatefix(core): mark .buf and .no_dma_buffers sections as NOLOADby cepetr · bce462ee · Sep 2, 2026 · 20 filesMessage 62 · AdequateLow 44Details
Commit message · cepetr
fix(core): mark .buf and .no_dma_buffers sections as NOLOAD
[no changelog]
62/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope! No meaningful explanatory body
Why it was queued
boot or update path
AI analysis · Low 44/100
This commit changes how the Trezor firmware's embedded linker scripts reserve memory for special buffer sections. It marks sections named .buf and .no_dma_buffers as NOLOAD, meaning the linker will reserve RAM space for them but will not include their initial contents in the firmware binary or copy them during startup. Without NOLOAD, these uninitialized buffers could be included in the binary image, increasing its size and potentially causing startup initialization issues. The change is a correctness fix for embedded memory layout, but the commit message does not frame it as a security fix and no exploit is demonstrated.
refactor(core/bootloader): drop the unreachable turn_on command list
The command enumeration has been dead since ad3cb11056 ("start the device on short press"). That commit widened the condition below it from `manufacturing_mode && cmd != BOOT_COMMAND_POWER_OFF` to just the latter, which subsumes the list entirely -- nothing the list selects can still change the outcome.
The widening was deliberate. What the list was trying to express is already covered: BOOT_COMMAND_POWER_OFF is set in exactly one place (reboot_and_power_off), so it alone marks the state that has no reason to boot. Every other command boots through -- and a boot with the button held still reaches the press loop via the button_is_down check just below, so long-press bootloader entry is unaffected.
Keeping the enumeration around only implies it still selects something, and any command added to it is a silent no-op.
[no changelog]
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
85/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Provides detailed explanatory context
Why it was queued
update trustboot or update path
AI analysis · Informational 15/100
This is a code cleanup in the Trezor bootloader. It removes a list of commands that was no longer doing anything useful, because a simpler check already covered all cases. There is no security bug being fixed here; the change only makes the code easier to understand and prevents future mistakes where someone might add a command to the now-deleted list and expect it to matter.
✓ Descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Provides an explanatory body✓ Links an issue, advisory, or supporting reference✓ Names security-relevant behavior explicitly
Why it was queued
memory safetyboot or update path
AI analysis · Low 44/100
This commit changes how a memory buffer section named .buf is handled during the build process for the Trezor Model T (T2T1) hardware wallet. By marking it as NOLOAD, the linker no longer includes this uninitialized buffer in the firmware image file, which prevents the firmware from exceeding its allowed flash storage region. This is a build/linker fix that avoids a firmware size overflow; it does not by itself appear to be an exploitable vulnerability, but overflowing a flash region could in some cases cause build failures, boot issues, or undefined behavior if an oversized image were written to the device.
✓ Descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope! No meaningful explanatory body
Why it was queued
boot or update path
AI analysis · Informational 15/100
This commit is a build-system feature change. It lets developers use the Address Sanitizer (ASAN) debugging tool with all Trezor firmware projects, not just the emulator. It also adds a guard that stops ASAN from being used in real hardware builds, where it is unsupported. There is no security vulnerability here.
✓ Descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Mentions testing or verification! No meaningful explanatory body
Why it was queued
cryptography-sensitive path
AI analysis · Informational 15/100
This commit adds automated tests for the ECDSA cryptographic code in the Trezor firmware. It also introduces a small, safer wrapper type (RecId) for signature recovery IDs so only valid values (0-3) can be passed to the underlying C code. There is no fix for a known vulnerability here; it is a hardening and test-coverage change.
refactor(core/rust/crypto: introduce testing harness and zkp support
moves the test-only `rand` insecure PRNG into a dedicated test-only module `testutil`
implements `random_buffer` extern "C" export that delegates to the insecure PRNG and applies a xor with a masking string <PRNG-Rust-Tests> so that its presence can be detected within a compiled artifact (required for secp256-zkp EC arithmetic even for verification because curve_to_jacobian needs random data for FIH masking)
explicitly runs the test harness and initializes the zkp context before start (also required by secp256-zkp)
moves the secp256-zkp default error handlers from modtrezorcrypto.c to the crypto crate, so that they're properly set for any user of crypto, not just firmware. (this removes the "raise ValueError" feature for illegal arguments -- that shouldn't matter too much, these are checks on "is the function called properly", not on validity of user provided input)
100/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context✓ Explains rationale or failure mode✓ Mentions testing or verification✓ Names security-relevant behavior explicitly
Why it was queued
entropy or randomnessdefensive validationcryptography-sensitive path
AI analysis · Informational 18/100
This commit reorganizes how a test-only random number generator and certain Bitcoin-related cryptographic error handlers are wired up in the Trezor firmware's Rust crypto code. It moves an insecure PRNG used only in unit tests into its own module and exports it under a detectable name so test builds can find it. It also moves default error callbacks for the secp256k1-zkp library from the Python module layer into the core crypto crate so all users of the crate share them. The change removes a Python ValueError raise for illegal library arguments, replacing it with a system error shutdown. There is no direct evidence in the commit of a vulnerability being fixed; it reads as a refactoring to support testing and a new cryptographic feature (zero-knowledge proofs).
Necessitated a refactor of previous CSlice code to use a shared internal helper. I also took the opportunity to change as_slice / as_slice_mut to always return a slice that may be empty, collapsing the NULL/empty distinction on the Rust side too.
Fun had been had on all sides, esp. with obscure provenance bugs and the correct safety properties.
80/100 · StrongMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Provides detailed explanatory context
Why it was queued
boot or update path
AI analysis · Low 29/100
This commit is a code-quality refactor in the Trezor firmware's Rust layer. It introduces a new mutable-slice helper (CSliceMut) and changes the existing CSlice so that empty/NULL slices return an empty Rust slice instead of an optional value that callers had to unwrap. The change is described by the developer as fixing 'obscure provenance bugs' and aligning safety properties. There is no direct evidence in the commit of an exploitable vulnerability; it reads as defensive hardening of the boundary between C and Rust code.
Add doxygen comments to `storage.h` in the style used across core, covering all public functions and the UI callback.
The NULL contract of every pointer argument is stated explicitly, matched against the syscall and smcall verifiers so that the header and the privilege boundary agree on which arguments are optional.
The only deliberate divergence is `storage_get()`, whose `val` ignores `max_len` when NULL, while the verifiers require `max_len` to be 0 in that case. The header documents the stricter rule.
Comments only, no functional change.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
[no changelog]
92/100 · StrongMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Provides detailed explanatory context✓ Explains rationale or failure mode
Why it was queued
credential or privilege state
AI analysis · Informational 15/100
This commit only adds documentation comments to a C header file. No program code was changed, so it cannot introduce or fix a security bug by itself. The comments describe how the storage API is already supposed to behave, including rules about which pointer arguments may be NULL.
✓ Descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope! No meaningful explanatory body
Why it was queued
cryptography-sensitive pathboot or update path
AI analysis · Low 35/100
This commit tightens a safety check in the Trezor hardware wallet's user-interface code. It makes sure that a small on-screen menu component can hold at least as many items as the firmware might ever ask it to display, preventing a capacity mismatch that could crash the device when a user is choosing recovery-seed word counts or similar options. There is no direct evidence in the commit that this crash was exploitable to steal funds or bypass security.
refactor(core): separate norcow config from flash layout
`norcow_config.h` lived in the flash module and was pulled in by `sys/flash.h`, while storage's `norcow.h` includes `norcow_config.h`, which itself includes `sys/flash.h` — a cycle between the two modules.
Split it along its natural seam:
- `STORAGE_AREAS` / `STORAGE_AREAS_COUNT` are flash layout (defined by `flash_layout.c`, used by `flash_utils.c` and the bootloader emulator), so they join the other area declarations in `sys/flash.h`. - The norcow parameters move to `sec/storage/inc/norcow_config.h`, next to the module that needs them, and `NORCOW_SECTOR_COUNT` is now derived from `STORAGE_AREAS_COUNT` so the two cannot disagree.
`projects/unix/main_main.c` uses `MODEL_INTERNAL_NAME` but reached `trezor_model.h` only through the removed include chain, so it now includes it directly.
Legacy and `storage/tests` are unaffected: each supplies its own `norcow_config.h`.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
[no changelog]
95/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Provides detailed explanatory context✓ Mentions testing or verification
Why it was queued
update trust
AI analysis · Informational 15/100
This commit is a straightforward internal code cleanup. It moves two configuration constants between header files to remove a circular include dependency between the flash and storage modules. There is no user-facing change, no bug fix, and no security-sensitive behavior change.
fix(core/rust): adjust menu item limits to prevent overflow errors
[no changelog]
79/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Explains rationale or failure mode✓ Names security-relevant behavior explicitly! No meaningful explanatory body
Why it was queued
memory safetyboot or update path
AI analysis · Low 42/100
This commit fixes a capacity mismatch in the Trezor hardware wallet's on-screen menus. A global limit allowed up to 6 menu entries, but the internal storage for two menu layouts only held 5. When a 6-item menu was built, the code would panic (crash) instead of showing the menu. The patch raises both storage capacities to 6 and adds compile-time checks so the bug cannot silently recur if the global limit is changed again. The crash is a denial-of-service issue for the device UI, not a direct theft-of-funds vulnerability.
Security candidatechore: use ruff instead of black and isortby obrusvit · 724737f5 · Aug 31, 2026 · 136 filesMessage 72 · AdequateInformational 15Details
Commit message · obrusvit
chore: use ruff instead of black and isort
- reducing the number of tools in the repo - in the future, we may replace also pylint and pyright
[no changelog]
72/100 · AdequateMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Provides an explanatory body
Why it was queued
cryptography-sensitive pathseed or entropy pathsigning or wallet pathboot or update pathauthentication pathparser or protocol path
AI analysis · Informational 15/100
This commit is a routine housekeeping change that swaps the Python code-formatting tools used in the project from `black` and `isort` to `ruff`. It updates configuration files, build scripts, and reformats many source files to match the new tool's style. There is no change to what the software actually does, no new features, and no indication of any security problem or fix.
✓ Descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope! No meaningful explanatory body
Why it was queued
cryptography-sensitive pathboot or update path
AI analysis · Informational 17/100
This commit is a code cleanup in the Trezor firmware's user-interface layer. It moves a small 'menu item intent' type to a better location, fixes a minor error-handling inconsistency, and reorders some on-screen menu entries so the cancel option appears first. There is no direct evidence this fixes an exploitable security vulnerability, but the reordering could be a defensive usability improvement to prevent accidental destructive actions.
Security candidatefix(core): refuse an emulator BLE read that cannot take a whole packetby tychovrahe · 6999d093 · Aug 31, 2026 · 1 fileMessage 85 · StrongLow 45Details
Commit message · tychovrahe
fix(core): refuse an emulator BLE read that cannot take a whole packet
The socket is SOCK_DGRAM, so `recvfrom()` consumes the datagram whole and discards whatever does not fit the buffer it is given. Passing `MIN(max_len, sizeof(buf))` therefore let a caller with a short buffer destroy part of a received packet and get only its own length back - the emulator's version of the defect just fixed in the hardware driver, reached through UDP truncation rather than a queue dequeue.
Reject a `max_len` below the packet size before reading, as the hardware driver now does, and read a whole buffer unconditionally: with the check above it, the MIN could only ever pick `sizeof(buf)`.
No caller is affected. `py_iface_read()` requires a buffer at least `RX_PACKET_SIZE` long and the bootloader wire layer requires an exact size; only a direct `ble_read` syscall could ask for less.
[no changelog]
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
85/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Provides detailed explanatory context
Why it was queued
update trust
AI analysis · Low 45/100
This commit fixes a bug in the Trezor emulator's Bluetooth-over-UDP code. Previously, if a caller asked to read a Bluetooth packet into a buffer smaller than a full packet, the emulator would silently throw away the leftover data. Now it refuses short reads up front, matching a fix already made in the real hardware driver. The commit message says no real callers were affected because all normal callers use full-size buffers.
Security candidatefix(core): stop ble_read() consuming a packet it cannot returnby tychovrahe · f9c35369 · Aug 31, 2026 · 1 fileMessage 95 · StrongLow 45Details
Commit message · tychovrahe
fix(core): stop ble_read() consuming a packet it cannot return
`ble_read()` dequeued from the RX queue and only then checked that the caller's buffer was big enough. `tsqueue_dequeue()` resets the entry and advances the read index unconditionally, and there is no re-enqueue on the error path, so a buffer below `BLE_RX_PACKET_SIZE` silently destroyed a received packet and reported zero bytes - indistinguishable to the caller from an empty queue.
The size test never needed the dequeued data. Once `read_len` is known to be `BLE_DATA_SIZE`, `max_len < read_len - BLE_DATA_HEADER_SIZE` is just `max_len < BLE_RX_PACKET_SIZE`, so hoist it above the dequeue. That also retires an expression that was only in bounds thanks to the `||` short-circuit ordering.
`py_iface_read()` rejects a short buffer before it gets here and the bootloader wire layer requires an exact size, so the only way in is an applet calling the `ble_read` syscall directly - which loses packets addressed to itself. No caller changes behaviour: for every `max_len` at or above the packet size the function does exactly what it did.
[no changelog]
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
95/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Provides detailed explanatory context✓ Mentions testing or verification
Why it was queued
update trust
AI analysis · Low 45/100
This commit fixes a bug in the Bluetooth code of Trezor hardware wallets. When a program asked to read a Bluetooth packet using a buffer that was too small, the code would first remove the packet from the receive queue and only then realize the buffer was too small. Because the removed packet was not put back, it was silently lost. The fix checks the buffer size before removing the packet, so an undersized read now simply reports 'nothing available' instead of destroying data. The commit notes that the only realistic way to trigger this is an applet making a direct system call with a too-small buffer, which would cause that applet to lose packets meant for it.
Security candidatefix(core): tie ble::read()'s length to its bufferby tychovrahe · 8b6adf43 · Aug 31, 2026 · 2 filesMessage 85 · StrongModerate 58Details
Commit message · tychovrahe
fix(core): tie ble::read()'s length to its buffer
`read()` took the buffer and the length the caller claims for it as separate arguments, and passed the length straight to `ble_read()`. `ble_read()` writes a full `BLE_RX_PACKET_SIZE` packet whenever the length it is given is at least that big, so a caller passing a length larger than its buffer gets 244 bytes written into something shorter.
The one caller, `py_iface_read()`, hands `read()` a Python-supplied buffer of arbitrary size and passed the constant `RX_PACKET_SIZE`. It checks `buf.len() < RX_PACKET_SIZE` first, so nothing is wrong today - but that check is the only thing standing between a short bytearray and an overflow.
Derive the length from the buffer so the two cannot disagree. `ble_read()` then refuses to write at all when the buffer is too small, and the check in `py_iface_read()` is left to produce a clearer error rather than to keep us in bounds. Saturate rather than fail the conversion: the length is only compared against the packet size, so a buffer above 65535 bytes must keep working.
[no changelog]
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
85/100 · StrongMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Provides detailed explanatory context✓ Names security-relevant behavior explicitly
Why it was queued
memory safety
AI analysis · Moderate 58/100
This commit fixes a potential buffer overflow in the Bluetooth Low Energy (BLE) read function of the Trezor hardware wallet firmware. Previously, the function trusted the caller's claim about how large the buffer was, which could allow more data to be written into a smaller memory area than intended. The fix makes the function use the actual buffer size instead, removing that trust relationship and preventing a possible overflow.
Security candidatefix(core): validate length in the nRF service callbacksby tychovrahe · a3fd85a0 · Aug 31, 2026 · 2 filesMessage 85 · StrongModerate 62Details
Commit message · tychovrahe
fix(core): validate length in the nRF service callbacks
`nrf_management_rx_cb()` and `nrf_test_cb()` select on `data[0]` without checking that a byte is present, and the two management handlers went further on the same assumption, computing `len - 1` and clamping the result with MIN(). A zero-length frame turned that subtraction into 4294967295, so the clamp always picked the full struct size.
The reads stayed inside the SPI frame buffer, so nothing was out of bounds. The defect is that a truncated response was accepted as a complete one:
- a one-byte MGMT_RESP_INFO set `info_valid` while leaving `drv->info` holding whatever a previous exchange left there, - a one-byte MGMT_RESP_AUTH_RESPONSE set `auth_data_valid` with `auth_data` still zeroed, which `nrf_authenticate()` then fed to the MAC comparison.
Both fail closed, but `nrf_authenticate()` gates boot in production builds and has no business deciding anything from a frame the nRF never filled in.
Require the whole payload before marking the state valid, and set the flag after the copy rather than before. The minimums match what the nRF sends: `send_info()` transmits `1 + sizeof(nrf_info_t)` bytes and `mgmt_process_challenge()` more than the digest we read. An older nRF that sent less would now fail `nrf_get_info()` and fall through to the SMP path, which forces an update - the outcome we want.
This is the same defect class as the recent fix to `ble_process_rx_msg_mac()`; these were the remaining callbacks.
[no changelog]
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
85/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Provides detailed explanatory context
Why it was queued
memory safetydefensive validation
AI analysis · Moderate 62/100
This commit fixes a bug in the Trezor hardware wallet's code that talks to the nRF wireless chip over SPI. The code read the first byte of a received message to decide what kind of response it was, without first checking whether any bytes had actually been received. For two important response types—device info and authentication data—it then marked the data as 'valid' before copying it. A zero-length or truncated message could therefore make the device believe it had received valid authentication data when it had not. The device 'fails closed' (it does not unlock anything from bad data), but the bug could still let an empty or truncated frame influence the boot-time authentication check, which is not a safe design. The fix requires the full expected payload length before marking anything valid and moves the 'valid' flag to after the copy.
Security candidatefix(core/fido2): use a static counter when returning a bogus signatureby Roman Zeyde · bd8c3eed · Aug 31, 2026 · 1 fileMessage 67 · AdequateLow 35Details
Commit message · Roman Zeyde
fix(core/fido2): use a static counter when returning a bogus signature
[no changelog]
67/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Names security-relevant behavior explicitly! No meaningful explanatory body
Why it was queued
signing boundaryauthentication path
AI analysis · Low 35/100
This commit changes how the Trezor hardware wallet handles FIDO2/WebAuthn 'silent authentication' attempts. When a website asks for authentication without user confirmation (user presence=false), Trezor now returns a fake signature and a fixed counter value of 0, instead of advancing the real signature counter. Previously, the code advanced the real counter even when returning a fake signature. The change prevents the real counter from leaking or being consumed during silent authentication attempts, which is a minor privacy and state-consistency improvement.