BS
← All projectsBlockstream

Blockstream Jade

Firmware for Jade and Jade Plus Bitcoin hardware wallets.

BitcoinHardware walletsElevated watch
Repository coverage

502 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.

106security candidates69second-pass queue437AI analyses
25commits · 30 days
67commits · 60 days
220commits · 180 days
417commits · 365 days
Backfill bands
Aug 5 → Feb 6216 seen38 candidatesComplete
Feb 6 → Jun 6131 seen9 candidatesComplete
Jun 6 → Jul 657 seen10 candidatesComplete
Jul 6 → Aug 510 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.

57/100 average clarity
32Strong · 80–100
182Adequate · 60–79
250Thin · 40–59
38Opaque · 0–39
5security 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.
Jon Griffiths31357282157
Mike Tolkachev431438152
Daniel Newton771668055
Mihailo Milenkovic878057
roderico333060
Valerio Vaccaro1526054
Jamie C. Driver525053
Gustavo Cateim917078
Austin-Fulbright414060
quantpoet111072
cmiguelrn111045
tokyovyse111063
Analysis record

Published AI watches

Last scanned 29 minutes ago

Informational 20 AI analysisMessage 65 · Adequate
BS BlockstreamBlockstream Jade BitcoinHardware wallets

fmt: fix and prevent future use of unsupported format spec "zu"

This commit fixes the use of the "%zu" printf format specifier, which is not supported by the ESP32 toolchain used in this project. The unsupported format could cause log messages and user-facing strings to display incorrect or garbage val…

Use of unsupported printf format specifier %zu on embedded targetPotential for malformed log/UI output due to undefined format behaviorAddition of CI guard to prevent recurrence
26042182by Jon Griffiths+21−145 files
No security note in commit
Low 33 AI analysisMessage 30 · Opaque
BS BlockstreamBlockstream Jade BitcoinHardware wallets

otp: release md context on failure and clean temporary buffers

This commit fixes cleanup in the one-time-password (OTP) code. Previously, if setting up the cryptographic hash context failed, the code could return without freeing that context, and temporary buffers holding sensitive counter/HMAC data w…

Resource leak fix: mbedtls_md_free() now called on prepare_md_ctx failure pathSensitive buffer zeroization: wally_bzero() added for rcnt_buf and hmacControl-flow change from macro-based early return to explicit success checks
6e2cb6b6by Mike Tolkachev+26−141 file
No security note in commit
Low 47 AI analysisMessage 25 · Opaque
BS BlockstreamBlockstream Jade BitcoinHardware wallets

identity: clean temporary hash buffer

This commit fixes a small but real security hygiene issue in Blockstream Jade, a hardware wallet. A temporary buffer that holds a sensitive cryptographic hash (HMAC-SHA512 output used to derive child private keys) was not marked as sensiti…

Sensitive memory not marked for clearing (missing SENSITIVE_PUSH/POP)Buffer holds HMAC-SHA512 output used in BIP32 hardened child key derivationPatch adds explicit sensitive-memory lifecycle management
b5c35779by Mike Tolkachev+2−01 file
No security note in commit
Low 36 AI analysisMessage 15 · Opaque
BS BlockstreamBlockstream Jade BitcoinHardware wallets

qrmode: clean temporary buffer

This commit adds cleanup of a temporary memory buffer used when turning data into QR codes on the Blockstream Jade hardware wallet. The buffer could previously remain in memory after use, potentially leaving sensitive QR data behind for la…

Sensitive temporary buffer not cleared before fixUse of SENSITIVE_PUSH/SENSITIVE_POP memory-clearing macrosQR payload data may include private or confidential material
34f68c20by Mike Tolkachev+2−01 file
No security note in commit
Low 43 AI analysisMessage 50 · Thin
BS BlockstreamBlockstream Jade BitcoinHardware wallets

pinserver: retain unit private key on url change

This commit changes how Blockstream Jade handles its PIN server settings. Previously, changing the PIN server web address (URL) also wiped the device's private PIN encryption key, forcing the device to create a fresh one. The patch keeps t…

PIN private key lifecycle change: key retained on URL-only pinserver updatePublic-key change still triggers private-key erasure/regenerationDebug-clean and full reset still erase private key
ddcd93a3by Mike Tolkachev+38−324 files
No security note in commit
Informational 15 AI analysisMessage 45 · Thin
BS BlockstreamBlockstream Jade BitcoinHardware wallets

add more boards and descriptions

This commit is a documentation update for the do-it-yourself (DIY) hardware section of the Blockstream Jade project. It adds descriptions and photos for additional supported circuit boards and updates image file references. There are no co…

f4fb150dby Valerio Vaccaro+65−1625 files
No security note in commit
Moderate 63 AI analysisMessage 50 · Thin
BS BlockstreamBlockstream Jade BitcoinHardware wallets

process: limit array sizes in sign_tx, register_multisig and sign_bip85_digests

This commit adds size limits to several message-handling functions in the Blockstream Jade hardware wallet firmware. It caps how many asset records, transaction inputs/outputs, multisig signers, and BIP85 RSA digests a single RPC message c…

Adds explicit upper bounds on attacker-controllable array sizesRejects zero-input/zero-output transactions before further processingMoves BIP85 digest count validation before memory allocation
e87f26bfby Daniel Newton+251−8114 files
No security note in commit
Informational 15 AI analysisMessage 28 · Opaque
BS BlockstreamBlockstream Jade BitcoinHardware wallets

Update CHANGELOG.md

This commit only updates the project's CHANGELOG.md file to add release notes for version 1.0.41. It does not change any source code, build scripts, or configuration. The changelog text itself states that version 1.0.41 contains important …

Changelog describes prior security fixes in release 1.0.41Mentions external security researchers and tools (Loupe, Kvazar)References security-relevant improvements: stack protection, entropy, secure version, memory scrubbing, thread safety
dd45a7dcby Jon Griffiths+49−01 file
Vendor flagged security relevance
Informational 15 AI analysisMessage 35 · Opaque
BS BlockstreamBlockstream Jade BitcoinHardware wallets

jadepy: update jadepy release

This commit simply bumps the version of a Python dependency called jade-client from 1.0.40 to 1.0.41, along with its matching cryptographic hashes and a fresh PGP signature. There is no code change visible here, and nothing in the commit m…

75eeada4by Jon Griffiths+19−192 files
No security note in commit
Informational 15 AI analysisMessage 45 · Thin
BS BlockstreamBlockstream Jade BitcoinHardware wallets

consistency: fix docs/comment/script typos

This commit fixes five spelling mistakes in documentation, code comments, and a script comment. None of the changes alter program behavior or affect security.

9ea7eed3by Jon Griffiths+5−55 files
No security note in commit
Informational 17 AI analysisMessage 50 · Thin
BS BlockstreamBlockstream Jade BitcoinHardware wallets

esp_lcd: vendor the esp_lcd component (ESP-IDF v5.5.4)

This commit copies the ESP-IDF 'esp_lcd' display-driver component directly into the Blockstream Jade firmware repository (a process called 'vendoring') at version 5.5.4. It adds about 8,700 lines of display-interface code for LCD panels us…

Large vendored third-party driver import (+8,718 lines, 43 files)Driver code runs in interrupt context and uses DMA, which historically carries memory-safety and cache-coherency risksNo explicit security fix or bug-fix narrative in commit title/message
d89207bfby Daniel Newton+8718−043 files
No security note in commit
Informational 21 AI analysisMessage 57 · Thin
BS BlockstreamBlockstream Jade BitcoinHardware wallets

build: "fakeprod" prod-like development mode

This commit adds an official developer-only 'fakeprod' build mode for Blockstream Jade hardware wallets. It deliberately weakens several security settings (single signature, committed private keys, readable flash-encryption key, enabled JT…

In-tree private keys committed to repository (tools/fakeprod_v2.pem, tools/fakeprod_attest.pem)Secure boot minimum signatures reduced from 2 to 1Flash encryption set to DEVELOPMENT mode with key readable
cecedc76by Daniel Newton+350−59 files
Vendor flagged security relevance
Low 45 AI analysisMessage 45 · Thin
BS BlockstreamBlockstream Jade BitcoinHardware wallets

urldecode: add validation for URL encoding

This commit hardens a URL-decoding helper in Blockstream Jade's firmware. It adds validation so malformed percent-encoding, embedded null bytes, control characters, and non-ASCII bytes are rejected instead of being silently decoded. A new …

Input validation added for percent-encoded sequencesRejection of embedded NUL bytes and control characters in decoded outputRejection of non-ASCII decoded bytes (>=0x80 and DEL 0x7f)
dd0d699dby Mike Tolkachev+100−272 files
No security note in commit
Low 26 AI analysisMessage 73 · Adequate
BS BlockstreamBlockstream Jade BitcoinHardware wallets

qr: split datastream data into its own allocation, reorder alloc/free calls

This change rewrites how a small QR-code scanning library allocates memory. Instead of one big 16 KB block, it now uses two separate 8 KB blocks, and it frees them in the reverse order they were allocated. The stated goal is to avoid memor…

Memory allocation pattern changed from single large allocation to two smaller allocationsAllocation/free ordering changed to LIFO-style reverse pairingPointer field added to struct that previously held an inline array
43640f8cby Jon Griffiths+24−153 files
No security note in commit
Low 46 AI analysisMessage 45 · Thin
BS BlockstreamBlockstream Jade BitcoinHardware wallets

psbt: check output amount presence before reading

This commit tightens how Blockstream Jade reads output amounts from a Partially Signed Bitcoin Transaction (PSBT). Before, the code tried to read the amount directly without first confirming it was actually present. Now it checks whether t…

Missing-field check added before sensitive value readPSBT output amount parsing hardenedDefensive validation of libwally presence flags
9f2b916dby Jon Griffiths+2−11 file
No security note in commit
Moderate 59 AI analysisMessage 50 · Thin
BS BlockstreamBlockstream Jade BitcoinHardware wallets

psbt: require that all psbt bytes are consumed when parsing

This commit tightens how Blockstream Jade verifies a Bitcoin transaction file (PSBT) before signing. It now requires that every byte in the file is part of the transaction, with no leftover or trailing data allowed. Previously, extra bytes…

Parsing now rejects trailing/unconsumed bytes in PSBT inputChange is in the signing code path, directly affecting what the hardware wallet will signUse of stricter libwally-core parse flag (COMPLETE) indicates parser hardening
605d86acby Jon Griffiths+1−11 file
No security note in commit
Informational 15 AI analysisMessage 68 · Adequate
BS BlockstreamBlockstream Jade BitcoinHardware wallets

hardware: add Jade Plus design files

This commit reorganizes existing hardware design files for the original Jade device into a new subfolder and adds the open-source hardware design files for a new product, the Jade Plus. It does not change any software, firmware, or device …

5ad85661by bitcoinhelp+2573−122520 files
No security note in commit
Informational 15 AI analysisMessage 55 · Thin
BS BlockstreamBlockstream Jade BitcoinHardware wallets

tests: migrate low level tests

This commit only moves and reorganizes existing test code. It adds a new helper to check whether the current transport is a specific type, moves a shared wait() function into a common test package file, and migrates a large set of low-leve…

767006a0by Valerio Vaccaro+1580−54 files
No security note in commit
Low 27 AI analysisMessage 68 · Adequate
BS BlockstreamBlockstream Jade BitcoinHardware wallets

mnemonic: restore recovery phrase by word number

This commit adds a new way to restore a Bitcoin wallet on a Blockstream Jade hardware device: instead of typing each recovery word, users can enter each word's one-based BIP39 number (for example, '1' for 'abandon'). The change also refact…

New user-input path for seed recovery (word numbers instead of words)Refactoring of mnemonic input/validation helpersAddition of SENSITIVE_PUSH/POP for local mnemonic buffers
99927491by roderico+580−3183 files
No security note in commit
Informational 18 AI analysisMessage 68 · Adequate
BS BlockstreamBlockstream Jade BitcoinHardware wallets

ui: support bounded variable-length digit entry

This commit refactors the on-screen digit-entry keypad in Blockstream Jade so callers can ask for shorter number fields (for example 4 digits instead of 6) and optionally cap the entered value. The existing 6-digit PIN and index entry beha…

New bounded numeric input path may prevent out-of-range values in future callersRefactoring centralizes input validation rather than relying on callersNo buffer-size changes observed; digit array remains DIGIT_ENTRY_SIZE (6) elements
69627745by roderico+128−332 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.

Security candidatefmt: fix and prevent future use of unsupported format spec "zu"by Jon Griffiths · 26042182 · Sep 21, 2026 · 5 filesMessage 65 · AdequateInformational 20Details
Commit message · Jon Griffiths

fmt: fix and prevent future use of unsupported format spec "zu"

Reported-by: oroderico <oroderico@users.noreply.github.com>

65/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides an explanatory body
Why it was queued
signing or wallet pathauthentication path
AI analysis · Informational 20/100

This commit fixes the use of the "%zu" printf format specifier, which is not supported by the ESP32 toolchain used in this project. The unsupported format could cause log messages and user-facing strings to display incorrect or garbage values. The commit also adds a CI check to prevent future use of "%zu". There is no direct evidence in the commit that this caused a security vulnerability such as information disclosure or memory corruption, but it is a code-quality fix that removes undefined/unsupported formatting behavior.

Security candidateotp: release md context on failure and clean temporary buffersby Mike Tolkachev · 6e2cb6b6 · Sep 18, 2026 · 1 fileMessage 30 · OpaqueLow 33Details
Commit message · Mike Tolkachev

otp: release md context on failure and clean temporary buffers

30/100 · OpaqueMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component! No meaningful explanatory body! Contains work-in-progress language! Opaque security-relevant change
Why it was queued
authentication path
AI analysis · Low 33/100

This commit fixes cleanup in the one-time-password (OTP) code. Previously, if setting up the cryptographic hash context failed, the code could return without freeing that context, and temporary buffers holding sensitive counter/HMAC data were not always wiped. The change ensures the context is released on failure and wipes temporary buffers before returning. It is a defensive hardening fix rather than a clear exploitable vulnerability.

AI review queuedidentity: clean temporary hash bufferby Mike Tolkachev · b5c35779 · Sep 11, 2026 · 1 fileMessage 25 · OpaqueLow 47Details
Commit message · Mike Tolkachev

identity: clean temporary hash buffer

25/100 · OpaqueMessage clarity
✓ Descriptive subject✓ Names a concrete action or component! No meaningful explanatory body! Contains work-in-progress language
Why it was queued
second-pass: opaque commit message
AI analysis · Low 47/100

This commit fixes a small but real security hygiene issue in Blockstream Jade, a hardware wallet. A temporary buffer that holds a sensitive cryptographic hash (HMAC-SHA512 output used to derive child private keys) was not marked as sensitive memory. That means the buffer could be left uncleared in RAM after use, potentially leaking key material if an attacker could read device memory. The patch marks the buffer with SENSITIVE_PUSH so it is properly wiped with SENSITIVE_POP. It is a cleanup/fix, not a full exploit chain by itself.

AI review queuedqrmode: clean temporary bufferby Mike Tolkachev · 34f68c20 · Sep 10, 2026 · 1 fileMessage 15 · OpaqueLow 36Details
Commit message · Mike Tolkachev

qrmode: clean temporary buffer

15/100 · OpaqueMessage clarity
✓ Descriptive subject! No meaningful explanatory body! Contains work-in-progress language
Why it was queued
second-pass: opaque commit message
AI analysis · Low 36/100

This commit adds cleanup of a temporary memory buffer used when turning data into QR codes on the Blockstream Jade hardware wallet. The buffer could previously remain in memory after use, potentially leaving sensitive QR data behind for later code to read. The fix marks the buffer as sensitive and securely clears it after use.

Security candidatepinserver: retain unit private key on url changeby Mike Tolkachev · ddcd93a3 · Aug 27, 2026 · 4 filesMessage 50 · ThinLow 43Details
Commit message · Mike Tolkachev

pinserver: retain unit private key on url change

50/100 · ThinMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Names security-relevant behavior explicitly! No meaningful explanatory body
Why it was queued
secret or key materialboot or update pathauthentication path
AI analysis · Low 43/100

This commit changes how Blockstream Jade handles its PIN server settings. Previously, changing the PIN server web address (URL) also wiped the device's private PIN encryption key, forcing the device to create a fresh one. The patch keeps the existing private key when only the URL changes, while still erasing it when the PIN server's public key changes. This avoids unnecessary key rotation and reduces the risk that a legitimate URL-only update silently breaks PIN security by generating a new key.

AI review queuedadd more boards and descriptionsby Valerio Vaccaro · f4fb150d · Aug 27, 2026 · 25 filesMessage 45 · ThinInformational 15Details
Commit message · Valerio Vaccaro

add more boards and descriptions

45/100 · ThinMessage clarity
✓ Descriptive subject✓ Names a concrete action or component! No meaningful explanatory body
Why it was queued
second-pass: unusually broad change
AI analysis · Informational 15/100

This commit is a documentation update for the do-it-yourself (DIY) hardware section of the Blockstream Jade project. It adds descriptions and photos for additional supported circuit boards and updates image file references. There are no code changes that affect security.

AI review queuedprocess: limit array sizes in sign_tx, register_multisig and sign_bip85_digestsby Daniel Newton · e87f26bf · Aug 25, 2026 · 14 filesMessage 50 · ThinModerate 63Details
Commit message · Daniel Newton

process: limit array sizes in sign_tx, register_multisig and sign_bip85_digests

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 · Moderate 63/100

This commit adds size limits to several message-handling functions in the Blockstream Jade hardware wallet firmware. It caps how many asset records, transaction inputs/outputs, multisig signers, and BIP85 RSA digests a single RPC message can contain. The change appears to be a hardening patch: without these limits, an attacker or malformed host application could send an extremely large array and potentially exhaust device memory, cause long processing delays, or trigger memory corruption during signing and multisig registration. The patch also makes the array-length check happen earlier and more consistently across the code.

AI review queuedUpdate CHANGELOG.mdby Jon Griffiths · dd45a7dc · Aug 21, 2026 · 1 fileMessage 28 · OpaqueInformational 15Details
Commit message · Jon Griffiths

Update CHANGELOG.md

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

This commit only updates the project's CHANGELOG.md file to add release notes for version 1.0.41. It does not change any source code, build scripts, or configuration. The changelog text itself states that version 1.0.41 contains important security updates and thanks several security researchers, but this commit is purely documentation and does not implement any of those updates.

AI review queuedjadepy: update jadepy releaseby Jon Griffiths · 75eeada4 · Aug 21, 2026 · 2 filesMessage 35 · OpaqueInformational 15Details
Commit message · Jon Griffiths

jadepy: update jadepy release

35/100 · OpaqueMessage clarity
✓ Descriptive subject! No meaningful explanatory body
Why it was queued
second-pass: opaque commit message
AI analysis · Informational 15/100

This commit simply bumps the version of a Python dependency called jade-client from 1.0.40 to 1.0.41, along with its matching cryptographic hashes and a fresh PGP signature. There is no code change visible here, and nothing in the commit message or diff indicates a security fix or vulnerability.

Security candidateconsistency: fix docs/comment/script typosby Jon Griffiths · 9ea7eed3 · Aug 19, 2026 · 5 filesMessage 45 · ThinInformational 15Details
Commit message · Jon Griffiths

consistency: fix docs/comment/script typos

45/100 · ThinMessage clarity
✓ Descriptive subject✓ Names a concrete action or component! No meaningful explanatory body
Why it was queued
seed or entropy path
AI analysis · Informational 15/100

This commit fixes five spelling mistakes in documentation, code comments, and a script comment. None of the changes alter program behavior or affect security.

AI review queuedesp_lcd: vendor the esp_lcd component (ESP-IDF v5.5.4)by Daniel Newton · d89207bf · Aug 14, 2026 · 43 filesMessage 50 · ThinInformational 17Details
Commit message · Daniel Newton

esp_lcd: vendor the esp_lcd component (ESP-IDF v5.5.4)

50/100 · ThinMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component! No meaningful explanatory body
Why it was queued
second-pass: unusually broad change
AI analysis · Informational 17/100

This commit copies the ESP-IDF 'esp_lcd' display-driver component directly into the Blockstream Jade firmware repository (a process called 'vendoring') at version 5.5.4. It adds about 8,700 lines of display-interface code for LCD panels using I2C, SPI, I80, RGB, and MIPI-DSI buses. The change itself is a build/dependency packaging step, not a fix for a known security bug. There is no indication in the commit message or diff that this is intended to address a security issue, and no external security references are provided.

Security candidatebuild: "fakeprod" prod-like development modeby Daniel Newton · cecedc76 · Aug 13, 2026 · 9 filesMessage 57 · ThinInformational 21Details
Commit message · Daniel Newton

build: "fakeprod" prod-like development mode

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
boot or update path
AI analysis · Informational 21/100

This commit adds an official developer-only 'fakeprod' build mode for Blockstream Jade hardware wallets. It deliberately weakens several security settings (single signature, committed private keys, readable flash-encryption key, enabled JTAG and ROM download mode) so developers can easily flash and debug a production-like firmware. The commit and its documentation explicitly warn that this mode is NOT secure and must never be used on a real wallet. It is a development/testing feature, not a vulnerability in production firmware.

AI review queuedurldecode: add validation for URL encodingby Mike Tolkachev · dd0d699d · Aug 12, 2026 · 2 filesMessage 45 · ThinLow 45Details
Commit message · Mike Tolkachev

urldecode: add validation for URL encoding

45/100 · ThinMessage clarity
✓ Descriptive subject✓ Names a concrete action or component! No meaningful explanatory body
Why it was queued
parser or protocol pathsecond-pass: security-sensitive path
AI analysis · Low 45/100

This commit hardens a URL-decoding helper in Blockstream Jade's firmware. It adds validation so malformed percent-encoding, embedded null bytes, control characters, and non-ASCII bytes are rejected instead of being silently decoded. A new 'dry run' validation function is also exposed so callers can check an encoded string before decoding it. The change is defensive: it reduces the chance that an attacker could smuggle dangerous characters into URLs or trick downstream code with truncated or malformed strings.

AI review queuedqr: split datastream data into its own allocation, reorder alloc/free callsby Jon Griffiths · 43640f8c · Aug 10, 2026 · 3 filesMessage 73 · AdequateLow 26Details
Commit message · Jon Griffiths

qr: split datastream data into its own allocation, reorder alloc/free calls

Two 8kb allocations are more likely to succeed than a single 16kb
allocation, when dram is fragmented.

Also reorder alloc/free calls into matched reverse pairs to reduce fragmentation.

73/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context
Why it was queued
parser or protocol pathsecond-pass: security-sensitive path
AI analysis · Low 26/100

This change rewrites how a small QR-code scanning library allocates memory. Instead of one big 16 KB block, it now uses two separate 8 KB blocks, and it frees them in the reverse order they were allocated. The stated goal is to avoid memory-allocation failures on a small embedded device when memory is fragmented. The patch itself does not fix a known exploitable bug, but it touches memory allocation and pointer handling in code that processes untrusted QR images, so it has defensive-security relevance.

Security candidatepsbt: check output amount presence before readingby Jon Griffiths · 9f2b916d · Aug 9, 2026 · 1 fileMessage 45 · ThinLow 46Details
Commit message · Jon Griffiths

psbt: check output amount presence before reading

45/100 · ThinMessage clarity
✓ Descriptive subject✓ Names a concrete action or component! No meaningful explanatory body
Why it was queued
signing boundarysigning or wallet path
AI analysis · Low 46/100

This commit tightens how Blockstream Jade reads output amounts from a Partially Signed Bitcoin Transaction (PSBT). Before, the code tried to read the amount directly without first confirming it was actually present. Now it checks whether the amount exists before reading it. This is a defensive hardening change that could prevent incorrect handling of missing or malformed PSBT output data during transaction signing.

Security candidatepsbt: require that all psbt bytes are consumed when parsingby Jon Griffiths · 605d86ac · Aug 9, 2026 · 1 fileMessage 50 · ThinModerate 59Details
Commit message · Jon Griffiths

psbt: require that all psbt bytes are consumed when parsing

50/100 · ThinMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component! No meaningful explanatory body
Why it was queued
signing boundarysigning or wallet path
AI analysis · Moderate 59/100

This commit tightens how Blockstream Jade verifies a Bitcoin transaction file (PSBT) before signing. It now requires that every byte in the file is part of the transaction, with no leftover or trailing data allowed. Previously, extra bytes at the end could have been ignored, which might let a malicious or malformed PSBT hide unexpected content or influence how the device interprets what it is signing.

AI review queuedhardware: add Jade Plus design filesby bitcoinhelp · 5ad85661 · Aug 5, 2026 · 20 filesMessage 68 · AdequateInformational 15Details
Commit message · bitcoinhelp

hardware: add Jade Plus design files

Adds the open source hardware files for Jade Plus, and groups each board's files into its own subdirectory so the two sets aren't interleaved in one flat folder.

68/100 · AdequateMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context
Why it was queued
second-pass: unusually broad change
AI analysis · Informational 15/100

This commit reorganizes existing hardware design files for the original Jade device into a new subfolder and adds the open-source hardware design files for a new product, the Jade Plus. It does not change any software, firmware, or device behavior. There is no indication of a security fix or vulnerability being addressed.

Security candidatetests: migrate low level testsby Valerio Vaccaro · 767006a0 · Jul 24, 2026 · 4 filesMessage 55 · ThinInformational 15Details
Commit message · Valerio Vaccaro

tests: migrate low level tests

55/100 · ThinMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Mentions testing or verification! No meaningful explanatory body
Why it was queued
authentication path
AI analysis · Informational 15/100

This commit only moves and reorganizes existing test code. It adds a new helper to check whether the current transport is a specific type, moves a shared wait() function into a common test package file, and migrates a large set of low-level RPC tests into a new file. There are no changes to the actual Jade firmware, wallet logic, or production code, so this cannot introduce a security vulnerability in the shipped product.

Security candidatemnemonic: restore recovery phrase by word numberby roderico · 99927491 · Jul 18, 2026 · 3 filesMessage 68 · AdequateLow 27Details
Commit message · roderico

mnemonic: restore recovery phrase by word number

Allow advanced 12- and 24-word recovery to use one-based BIP39 word numbers. Resolve and confirm every word, retain the existing final-word calculation flow, and validate the complete recovery phrase before deriving keys. Clear sensitive entry state and displayed words when the flow exits.

68/100 · AdequateMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context
Why it was queued
secret or key materialdefensive validation
AI analysis · Low 27/100

This commit adds a new way to restore a Bitcoin wallet on a Blockstream Jade hardware device: instead of typing each recovery word, users can enter each word's one-based BIP39 number (for example, '1' for 'abandon'). The change also refactors the existing word-entry code and adds cleanup of sensitive on-screen state when the user exits. There is no direct evidence in the commit that this fixes an active security vulnerability, but it touches sensitive seed-recovery code and changes how user input is validated and cleared.

Security candidateui: support bounded variable-length digit entryby roderico · 69627745 · Jul 18, 2026 · 2 filesMessage 68 · AdequateInformational 18Details
Commit message · roderico

ui: support bounded variable-length digit entry

Allow digit-entry callers to select a shorter field width and an optional maximum value while preserving the existing six-digit PIN and index defaults. Skip unavailable candidate digits during navigation and keep short-number entry semantics unchanged.

68/100 · AdequateMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context
Why it was queued
access control
AI analysis · Informational 18/100

This commit refactors the on-screen digit-entry keypad in Blockstream Jade so callers can ask for shorter number fields (for example 4 digits instead of 6) and optionally cap the entered value. The existing 6-digit PIN and index entry behavior is kept unchanged. There is no direct evidence in the commit that this fixes an active security bug; it reads as a feature/generalization to support new UI flows such as entering BIP39 word numbers.

Security candidatetests: remove pin files for new tests, remove nolegacy from old testsby Jon Griffiths · 92380f9e · Jul 18, 2026 · 7 filesMessage 75 · AdequateInformational 15Details
Commit message · Jon Griffiths

tests: remove pin files for new tests, remove nolegacy from old tests

Note that for non-legacy testing we now only need to run the tx signing
tests that it affects.

75/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides an explanatory body✓ Mentions testing or verification
Why it was queued
signing boundaryaccess controlauthentication path
AI analysis · Informational 15/100

This commit is purely a testing infrastructure cleanup. It removes leftover command-line options and test files related to an older transaction-signing workflow, and makes sure temporary PIN files created during automated tests are cleaned up. There is no change to the actual Jade wallet firmware or its security behavior.

Lower-prioritywally: update wally to release 1.5.6by Jon Griffiths · cfce08ce · Jul 11, 2026 · 1 fileMessage 45 · ThinInformational 0Details
Commit message · Jon Griffiths

wally: update wally to release 1.5.6

45/100 · ThinMessage clarity
✓ Descriptive subject✓ Names a concrete action or component! No meaningful explanatory body
AI analysis · Informational 0/100

This commit updates a bundled dependency (libwally-core) to a newer release, but no actual diff content or release notes were provided. We cannot see what changed or whether any security issue is involved.

Lower-prioritypower: release PMU devices before deleting the i2c bus on M5CoreS3by Gustavo Cateim · fb3a009a · Jul 11, 2026 · 1 fileMessage 100 · StrongLow 26Details
Commit message · Gustavo Cateim

power: release PMU devices before deleting the i2c bus on M5CoreS3

Since ESP-IDF v5.5 i2c_del_master_bus() refuses to delete a bus that
still has devices attached and returns ESP_ERR_INVALID_STATE. On the
M5CoreS3, power_init() attaches the axp2101 and aw9523 devices and
then deletes the bus without releasing them, so under IDF 5.5.4 the
deinit fails, power_init() propagates the error and the JADE_ASSERT
in main.c aborts the boot. Same root cause as the touchscreen fix in
the previous commit (#306).

Remove both devices from the bus before deleting it.

Not tested on hardware (none available); the change mirrors the
touchscreen fix validated on the Waveshare S3 Touch LCD 2.

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✓ Links an issue, advisory, or supporting reference
Why it was queued
defensive validation
AI analysis · Low 26/100

This commit fixes a boot failure on the M5CoreS3 hardware. A recent update to the underlying Espressif SDK (ESP-IDF v5.5) changed the rules for deleting an I2C bus: it now refuses to delete a bus that still has devices attached. The power initialization code was attaching two power-management chips to the bus and then trying to delete the bus without detaching them first, causing the device to abort boot. The fix simply detaches those two devices before deleting the bus. There is no security vulnerability here—it's a compatibility/bug fix for a crash-on-boot scenario.

Security candidateinput: release touch handles before deleting the i2c busby Gustavo Cateim · 36776fec · Jul 11, 2026 · 1 fileMessage 86 · StrongLow 27Details
Commit message · Gustavo Cateim

input: release touch handles before deleting the i2c bus

Since ESP-IDF v5.5 (espressif/esp-idf@e73d78ba) i2c_del_master_bus()
refuses to delete a bus that still has devices attached and returns
ESP_ERR_INVALID_STATE. The touchscreen task registers a panel IO
device on the bus but never releases it, so under IDF 5.5.4 the first
touchscreen_deinit() fails and aborts via ESP_ERROR_CHECK.

On boards that panic with PRINT_HALT, such as the Waveshare S3 Touch
LCD 2, this freezes the device on the boot splash screen while
gathering camera entropy, and also hangs the QR scanner and the
camera exit path, since the touchscreen is restarted around every
camera start/stop (see #306).

Releasing the touch handle and the panel IO device (which removes it
from the bus) before deleting the bus restores the deinit/init cycle,
and also fixes a heap leak present under IDF 5.4 where both handles
leaked on every cycle.

Co-authored-by: oroderico <oroderico@users.noreply.github.com>

86/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context✓ Links an issue, advisory, or supporting reference✓ Names security-relevant behavior explicitly
Why it was queued
entropy or randomness
AI analysis · Low 27/100

This commit fixes a bug where the touchscreen driver failed to clean up internal handles when shutting down. On newer ESP-IDF firmware, this cleanup failure caused the device to panic and freeze during boot and when using the camera or QR scanner. The fix releases the touch and panel IO handles before deleting the I2C bus, preventing freezes and a memory leak. There is no evidence this is a security vulnerability or exploitable by an attacker.

Lower-prioritydiy: Add M5StickC Plus2 backlight brightness controlby oroderico · 5c22b7d3 · Jul 10, 2026 · 4 filesMessage 50 · ThinInformational 15Details
Commit message · oroderico

diy: Add M5StickC Plus2 backlight brightness control

50/100 · ThinMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component! No meaningful explanatory body
AI analysis · Informational 15/100

This commit adds a screen brightness control feature for the M5StickC Plus2 hardware variant of the Blockstream Jade device. It is a hardware-support and user-interface change, not a security fix or vulnerability. There is no indication of a security issue in the code changes.