ChipLab

Findings

Bugs ChipLab's AI agents found in open-source hardware, EDA tools and numerical libraries. One agent finds and documents each bug against the specification; a different agent reproduces it from scratch before it is reported to the project's maintainers.

Reports filed
214 190 public, 24 via security teams
Projects (public reports)
187
Fixed upstream
18

Each report carries a CLAB reference; this page links to it as findings.html#clab-2026-NNN. Security-relevant findings go through each project's private disclosure channels; they are counted above but listed here only once the fix is public. Updated 14 October 2026, 09:22 UTC.

RefProjectFindingStatusFiled
CLAB-2026-001olofk/servReading mie clears MTIE (csrr / csrrs / csrrc with x0 disables timer interrupts)
olofk/serv/issues/173		Open2026-10-08
(fix for 001)olofk/servserv_csr: Return MTIE when reading mie
olofk/serv/pull/174		Open2026-10-09
CLAB-2026-022Wren6991/Hazard3[alu] amominu.w / amomaxu.w store the signed min/max when Zbb is disabled (default A=1, Zbb=0)
Wren6991/Hazard3/issues/52		Open2026-10-09
CLAB-2026-023Wren6991/Hazard3[dm] A host write to data0 while an abstract command is busy changes data0 (spec: value unchanged)
Wren6991/Hazard3/issues/51		Open2026-10-09
CLAB-2026-024Wren6991/Hazard3[dm] sbcs.sbaccess resets to 0 (8-bit) instead of 2 (32-bit) when System Bus Access is enabled
Wren6991/Hazard3/issues/50		Open2026-10-09
CLAB-2026-026stnolting/neorv32[slink] RX_FULL / TX_EMPTY / TX_NFULL interrupt enables are at bits 18/19/21 in the RTL but 17/18/19 in the datasheet and neorv32_slink.h
stnolting/neorv32/issues/1662		Open2026-10-09
CLAB-2026-027stnolting/neorv32[dma] Byte-to-word transfer writes the byte replicated (0x34343434) instead of zero-extended (0x00000034)
stnolting/neorv32/issues/1663		Open2026-10-09
CLAB-2026-028antonblanchard/microwattFPU: fcfidus still misrounds some inputs above 2^53 (follow-up to #457)
antonblanchard/microwatt/issues/476		Open2026-10-10
CLAB-2026-036chipsalliance/Cores-VeeR-EH2minstret adds an extra 0x4000 at every low-byte carry (regression from the XLEN parameter change)
chipsalliance/Cores-VeeR-EH2/issues/56		Open2026-10-10
CLAB-2026-037verilator/verilatorNarrower signed case item is sign-extended against an unsigned case expression
verilator/verilator/issues/8720		Fixed2026-10-10
CLAB-2026-038darklife/darkriscvJALR does not clear bit 0 of the jump target
darklife/darkriscv/issues/114		Fixed2026-10-10
CLAB-2026-039chipfoundry/cf-sensor-soc-dsmSix functional issues found in RTL simulation of the sensor SoC digital ring (DSM and SAR)
chipfoundry/cf-sensor-soc-dsm/issues/1		Open2026-10-10
CLAB-2026-040chipfoundry/caravel-liteThree findings from RTL simulation of the Caravel harness (GPIO data writes, DLL output divider, GPIO36 staging)
chipfoundry/caravel-lite/issues/10		Open2026-10-10
CLAB-2026-041openhwgroup/cv-hpdcacheWrite buffer: threshold changed to 0 never expires an OPEN entry; wbuf_flush_i pulse does not hold back acceptance of later writes
openhwfoundation/cv-hpdcache/issues/156		Open2026-10-10
CLAB-2026-030enjoy-digital/litedram + liteeth
private email to [email protected]		Sent, awaiting reply2026-10-11
CLAB-2026-031pulp-platform/iDMA
private email to [email protected]		Sent, awaiting reply2026-10-11
CLAB-2026-032BMsemi/dw_openframe_user_project_FABulous_X2
private email to [email protected]		Sent, awaiting reply2026-10-11
CLAB-2026-033chipfoundry/caravel
private email to [email protected]		Sent, awaiting reply2026-10-11
CLAB-2026-034chipsalliance/rocket-chip-blocks
private email to [email protected]		Sent, awaiting reply2026-10-11
CLAB-2026-035chipsalliance/rocket-chip
private email to [email protected]		Sent, awaiting reply2026-10-11
CLAB-2026-042ucb-bar/gemminiOutput-stationary matmul with local C stride > 1 writes its result to the wrong scratchpad rows and overwrites other rows
ucb-bar/gemmini/issues/406		Open2026-10-11
CLAB-2026-044steveicarus/iverilogTwo differences from IEEE 1800-2017: constant $clog2 of a narrow signed value, and procedural copy between opposite-direction unpacked arrays
steveicarus/iverilog/issues/1574		Open2026-10-11
CLAB-2026-046pulp-platform/axi_riscv_atomics
private email to [email protected]		First contact sent2026-10-11
CLAB-2026-047pulp-platform/araSegment load/store: wrong field register group at EMUL>1, and a restarted (vstart>0) segment load skips field 0
pulp-platform/ara/issues/504		Open2026-10-11
CLAB-2026-049zachjs/sv2vSize cast removed when it bounds the width of a signed expression (`1'(-1'sb1)`)
zachjs/sv2v/issues/333		Open2026-10-11
CLAB-2026-056NVIDIA/Fuser`var` / `var_mean` over an expanded (stride-0) dimension return inf, nan or wrong values
NVIDIA/Fuser/issues/6052		Open2026-10-11
CLAB-2026-061ultraembedded/core_sdram_axi4AXI write beat with WSTRB=0 is never accepted; AS4C16M16S gets CAS 2 at 100 MHz and a 100 us startup delay
ultraembedded/core_sdram_axi4/issues/5		Open2026-10-11
CLAB-2026-062openhwfoundation/core-v-mcu[BUG] uDMA: I2C repeat count wraps above 128, quad LSB-first SPI receive shifted 3 bits, camera VSYNC_POLARITY reversed
openhwfoundation/core-v-mcu/issues/376		Open2026-10-11
CLAB-2026-072NVIDIA/cutlass[BUG] sm120 GEMM hangs when K = 0 (grouped and non-grouped) after can_implement(), initialize() and run() return success
NVIDIA/cutlass/issues/3733		Open2026-10-11
CLAB-2026-079chipsalliance/chiselShiftRegister.mem exposes undefined read data after stalls in tested SRAM configurations
chipsalliance/chisel/issues/5513		Open2026-10-11
CLAB-2026-080OpenXiangShan/CoupledL2TL2TL: advertised Put/atomics are acknowledged without updating memory, and sub-block Get/atomic responses are always size 6 with two beats
OpenXiangShan/CoupledL2/issues/530		Open2026-10-11
CLAB-2026-084YosysHQ/yosysarith_tree drops a bit permutation between chained adders
YosysHQ/yosys/issues/6324		Open2026-10-11
CLAB-2026-086llvm/circt[FIRRTL] SubOfPad drops the high bit of UInt subtraction
llvm/circt/issues/11276		Open2026-10-11
CLAB-2026-094B-Lang-org/bscFloatingPoint min and max select the opposite operand for distinct negative Float values
B-Lang-org/bsc/issues/1143		Open2026-10-11
CLAB-2026-101nvdla/hwCSB accepted pipelined reads return out of order and lose responses
nvdla/hw/issues/377		Open2026-10-11
CLAB-2026-107OpenPOWERFoundation/a2oCTR-decrementing branches test the full CTR in 32-bit mode
OpenPOWERFoundation/a2o/issues/25		Open2026-10-11
CLAB-2026-120alexforencich/cocotbext-axiAxiStreamSource tx_complete fires before final-beat acceptance
alexforencich/cocotbext-axi/issues/134		Open2026-10-11
CLAB-2026-124bespoke-silicon-group/basejump_stlbsg_mul_add_unsigned infers too few output bits for 16-bit and 32-bit operands
bespoke-silicon-group/basejump_stl/issues/757		Fixed2026-10-11
CLAB-2026-145xtensor-stack/xsimdBatch nearbyint and rint do not follow the current rounding mode
xtensor-stack/xsimd/issues/1424		Open2026-10-11
CLAB-2026-150ghdl/ghdlTEXTIO READ loses the sign of a negative integer with leading zeros
ghdl/ghdl/issues/3420		Fixed2026-10-11
CLAB-2026-160uxlfoundation/oneDNNx64 JIT binary-select post-op treats negative s8 conditions as false
uxlfoundation/oneDNN/issues/6225		Open2026-10-11
CLAB-2026-169ZipCPU/sdspiSTARTUP_CLOCKS=128 or 256 produces only one SPI startup clock
ZipCPU/sdspi/issues/28		Open2026-10-11
CLAB-2026-176hlslibs/ac_typesac_float::sub loses three quanta after an exact multiplication
hlslibs/ac_types/issues/39		Open2026-10-11
CLAB-2026-182Z3Prover/z3fp.roundToIntegral rounds the FP(2,3) subnormal 1/4 to 1 in nearest modes
Z3Prover/z3/issues/11139		Fixed2026-10-11
CLAB-2026-185ARM-software/CMSIS-NN[Bug]: arm_softmax_s16 doubles outputs at a normalization boundary and wraps one-class output
ARM-software/CMSIS-NN/issues/248		Open2026-10-11
CLAB-2026-187ARM-software/CMSIS-DSP[Bug]: arm_var_q15 wraps instead of applying the documented Q15 saturation
ARM-software/CMSIS-DSP/issues/354		Open2026-10-11
CLAB-2026-189SRI-CSL/yices2MCSat returns sat for opposite check-sat-assuming literals
SRI-CSL/yices2/issues/674		Open2026-10-11
CLAB-2026-191onnx/onnxReferenceEvaluator Resize rejects legacy configurations and returns incorrect nearest values
onnx/onnx/issues/8590		Open2026-10-11
CLAB-2026-213nickg/nvcREAL'VALUE repeats the first digit of multi-digit decimal exponents
nickg/nvc/issues/1687		Fixed2026-10-11
CLAB-2026-214amd/aocl-libm-oseA NaN lane makes vector erff return 6136.171875 for a finite input of 5
amd/aocl-libm-ose/issues/78		Open2026-10-11
CLAB-2026-219OpenMathLib/OpenBLASCROTG/ZROTG return incorrect rotations for scaled and unbalanced finite inputs
OpenMathLib/OpenBLAS/issues/6120		Fixed2026-10-11
CLAB-2026-241MikePopoloski/slangWide signed division and remainder fold to values with the wrong sign
MikePopoloski/slang/issues/2005		Open2026-10-11
CLAB-2026-247DrTimothyAldenDavis/GraphBLASA valued floating-point mask treats -0.0 as true
DrTimothyAldenDavis/GraphBLAS/issues/465		Open2026-10-11
CLAB-2026-256stp/stpNested let in a binding's right-hand side exposes an earlier sibling binding
stp/stp/issues/1318		Fixed2026-10-11
CLAB-2026-257apache/arrow[C++][Python] Safe decimal narrowing to Decimal32 drops coefficient bits before checking
apache/arrow/issues/53288		Open2026-10-11
CLAB-2026-270flame/blisHERK and HER2K propagate an ignored imaginary diagonal into the real result
flame/blis/issues/962		Open2026-10-11
CLAB-2026-275iree-org/ireeStableHLO stable sort reverses equal-key payloads on llvm-cpu
iree-org/iree/issues/25022		Open2026-10-11
CLAB-2026-284berkeley-abc/abcread_blif treats the clock token as init when latch initialization is omitted
berkeley-abc/abc/issues/587		Open2026-10-11
CLAB-2026-290chipsalliance/SurelogConstant-function += changes its RHS local during elaboration
chipsalliance/Surelog/issues/4214		Open2026-10-11
CLAB-2026-296OpenTimer/OpenTimerRising-to-falling setup check selects capture at 0.9 ns instead of 0.3 ns
OpenTimer/OpenTimer/issues/100		Open2026-10-11
CLAB-2026-317heitzmann/gdstkOASIS export writes full RobustPath width into the PATH half-width field
heitzmann/gdstk/issues/335		Open2026-10-11
CLAB-2026-331YosysHQ/eqyPublic builtin-cell matching returns PASS for unequal designs
YosysHQ/eqy/issues/103		Open2026-10-11
CLAB-2026-332UVVM/UVVMaxi_write waits for AWREADY before issuing later write-data beats
UVVM/UVVM/issues/305		Open2026-10-11
CLAB-2026-340OpenXiangShan/YunSuanFP16 vfrec7 returns incorrect estimates in the CVT32 lane
OpenXiangShan/YunSuan/issues/225		Open2026-10-11
CLAB-2026-341calyxir/calyxFixed-point dividers return remainder 77 for exact division of 100.5 by 0.5
calyxir/calyx/issues/2727		Open2026-10-11
CLAB-2026-353lsils/mockturtleVerilog reader imports NOR and XNOR assignments as constant zero
lsils/mockturtle/issues/712		Open2026-10-11
CLAB-2026-357myhdl/myhdlSigned floor division converts to truncation in Verilog and VHDL
myhdl/myhdl/issues/461		Open2026-10-11
CLAB-2026-362pymtl/pymtl3sext of a four-bit addition emits the wrong sign-bit expression
pymtl/pymtl3/issues/293		Open2026-10-11
CLAB-2026-379janestreet/hardcamlInitialized writable memory has two persistent drivers in generated VHDL
janestreet/hardcaml/issues/32		Open2026-10-11
CLAB-2026-384gtkwave/gtkwavevcd2lxt2 stores wrong settled values after multiple changes at one timestamp
gtkwave/gtkwave/issues/521		Fixed2026-10-11
CLAB-2026-385YosysHQ/prjtrellisecppack puts all frames in delta SVF with background, frequency or SPI options
YosysHQ/prjtrellis/issues/271		Open2026-10-11
CLAB-2026-390verilog-to-routing/vtr-verilog-to-routingParmys zero-extends signed operands in a widening addition
verilog-to-routing/vtr-verilog-to-routing/issues/3898		Open2026-10-11
CLAB-2026-401stanford-centaur/ponoBTOR2 usubo reports no borrow for 0 - 15 at width 4
stanford-centaur/pono/issues/689		Open2026-10-11
CLAB-2026-411arminbiere/aigeraigtoblif changes an explicit latch initialization of 1 to 0
arminbiere/aiger/issues/23		Open2026-10-11
CLAB-2026-419YosysHQ/apiculapROM packing omits later initialization when INIT_RAM_00 is absent
YosysHQ/apicula/issues/550		Open2026-10-11
CLAB-2026-426tukl-msd/DRAMPowerHBM2 refresh completion leaves active accounting intervals open
tukl-msd/DRAMPower/issues/159		Open2026-10-11
CLAB-2026-128black-parrot/black-parrotZba SHxADD instructions select rs2 as the shift amount
black-parrot/black-parrot/issues/1339		Open2026-10-12
CLAB-2026-233openvinotoolkit/openvino[Bug]: ONNX QuantizeLinear-13 shifts tie rounding with odd zero points
openvinotoolkit/openvino/issues/38755		Open2026-10-12
CLAB-2026-421veryl-lang/verylLocal concatenation in always_ff is emitted as a nonblocking assignment
veryl-lang/veryl/issues/3535		Open2026-10-12
CLAB-2026-436YosysHQ/prjpeppercorngmunpack rejects default gmpack output after the PLL padding change
YosysHQ/prjpeppercorn/issues/26		Open2026-10-12
CLAB-2026-448rggen/rggenSystemVerilog wrap counter subtracts inactive down inputs
rggen/rggen/issues/297		Fixed2026-10-12
CLAB-2026-452The-OpenROAD-Project/OpenROADDEF ROW without STEP imports zero site spacing
The-OpenROAD-Project/OpenROAD/issues/11706		Open2026-10-12
CLAB-2026-458StefanSchippers/xschemSPICE global bundle declaration keeps the range instead of the member nodes
StefanSchippers/xschem/issues/512		Open2026-10-12
CLAB-2026-463tukl-msd/DRAMSysStore-mode split transactions skip storage reads and writes
tukl-msd/DRAMSys/issues/72		Open2026-10-12
CLAB-2026-464alexforencich/cocotbext-ethGmiiSink drops the first valid octet and its error qualifier
alexforencich/cocotbext-eth/issues/10		Open2026-10-12
CLAB-2026-473Xilinx/brevitasShifted IntQuantTensor negation and subtraction return wrong real values
Xilinx/brevitas/issues/1636		Open2026-10-12
CLAB-2026-493Xyce/XyceB-source conflates V(V1) and I(V1) when the node and source share a name
Xyce/Xyce/issues/181		Open2026-10-12
CLAB-2026-510enjoy-digital/litejesd204N=14/N'=16 transport words place and recover conversion bits incorrectly
enjoy-digital/litejesd204/issues/10		Open2026-10-12
CLAB-2026-516fastmachinelearning/qonnxQCDQ import wraps asymmetric quantized weights before dequantization
fastmachinelearning/qonnx/issues/244		Open2026-10-12
CLAB-2026-542gatecat/prjoxideLRAM initialized data is lost during unpack and repack
gatecat/prjoxide/issues/54		Open2026-10-12
CLAB-2026-544VHDL-LS/rust_hdlLegal indexing of a function result is diagnosed as a function argument
VHDL-LS/rust_hdl/issues/518		Open2026-10-12
CLAB-2026-548huggingface/candleONNX Transpose with omitted perm swaps only the last two axes
huggingface/candle/issues/4034		Open2026-10-12
CLAB-2026-558halide/HalideInteger lerp with a float weight returns incorrectly rounded results
halide/Halide/issues/9546		Open2026-10-12
CLAB-2026-559scikit-rf/scikit-rfTouchstone 1.1 option line drops the second port reference resistance
scikit-rf/scikit-rf/issues/1450		Open2026-10-12
CLAB-2026-568pulp-platform/obi_peripheralsobi_uart: divisor 1 suppresses baud enables and stalls TX/RX
pulp-platform/obi_peripherals/issues/11		Open2026-10-12
CLAB-2026-569sonos/tractSigned integer Div by a power-of-two constant rounds negative quotients down
sonos/tract/issues/3003		Fixed2026-10-12
CLAB-2026-574arminbiere/cadicalimplied() exports a decision from a temporary disjunctive constraint
arminbiere/cadical/issues/205		Fixed2026-10-12
CLAB-2026-580hdl-modules/hdl-modulesaxi_lite_register_file: RDATA changes during a stalled status-register read
hdl-modules/hdl-modules/issues/231		Open2026-10-12
CLAB-2026-584ARM-software/optimized-routinesSVE pow/powf return zero or infinity for negative bases beside active zero or subnormal lanes
ARM-software/optimized-routines/issues/98		Fixed2026-10-12
CLAB-2026-586SanDisk-Open-Source/pyvcdCompound-vector aliases write a tuple in the VCD size field
SanDisk-Open-Source/pyvcd/issues/43		Open2026-10-12
CLAB-2026-588google/ruyPacked cache omits source layout and reuses the wrong operand for Mul(A, A)
google/ruy/issues/381		Open2026-10-12
CLAB-2026-589gem5/gem5RISC-V Zfa: finite fminm.s/fmaxm.s results are not NaN-boxed
gem5/gem5/issues/3599		Open2026-10-12
CLAB-2026-596riscv/riscv-opcodesGenerated Rust masks reject legal V segment loads and stores
riscv/riscv-opcodes/issues/439		Open2026-10-12
CLAB-2026-623microsoft/microxcalingFinite FP16 MX blocks become NaNs on the default PyTorch path
microsoft/microxcaling/issues/46		Open2026-10-12
CLAB-2026-635google/double-conversionALLOW_OCTALS accepts decimal suffixes after an all-zero integer prefix
google/double-conversion/issues/330		Open2026-10-12
CLAB-2026-645limbo018/DREAMPlaceDEF writer duplicates DO/BY/STEP when a ROW spans lines
limbo018/DREAMPlace/issues/244		Open2026-10-12
CLAB-2026-646flintlib/flintarb_pow loses containment when the output aliases an inexact exponent
flintlib/flint/issues/2879		Open2026-10-12
CLAB-2026-662microsoft/DirectXMathSix SIMD packed stores expose stale typed destination values with GCC at -O2
microsoft/DirectXMath/issues/355		Open2026-10-12
CLAB-2026-695fastfloat/fast_floatfrom_chars overwrites std::float32_t/std::float64_t on invalid input
fastfloat/fast_float/issues/424		Fixed2026-10-12
CLAB-2026-698abseil/abseil-cpp[Bug]: StrFormat omits zero padding in slow scientific %e and %g conversions
abseil/abseil-cpp/issues/2194		Open2026-10-12
CLAB-2026-705AcademySoftwareFoundation/Imath4D outerProduct uses the X component in two W-column entries
AcademySoftwareFoundation/Imath/issues/611		Open2026-10-12
CLAB-2026-716JuliaMath/openlibmnextafter and nexttoward return the finite endpoint in directed rounding modes
JuliaMath/openlibm/issues/361		Open2026-10-12
CLAB-2026-737apytypes/apytypesFixed-point cast rounds exact 0.75 to -1 instead of 1
apytypes/apytypes/issues/1043		Fixed2026-10-12
CLAB-2026-744boostorg/safe_numericsAutomatic promotion gives 231 for 999 % 1001
boostorg/safe_numerics/issues/152		Open2026-10-12
CLAB-2026-745HDFGroup/hdf5Generic float-to-int conversion maps in-range -0.25 to INT32_MIN
HDFGroup/hdf5/issues/6725		Open2026-10-12
CLAB-2026-756AcademySoftwareFoundation/OpenColorIOCPU simplification drops the final clamp of reverse ASC_CDL
AcademySoftwareFoundation/OpenColorIO/issues/2359		Open2026-10-12
CLAB-2026-759boostorg/jsonExact double -2^63 is rejected by conversion to std::int64_t
boostorg/json/issues/1209		Open2026-10-12
CLAB-2026-761diffblue/cbmcBit-field compound assignment converts RHS before multiplication
diffblue/cbmc/issues/9181		Open2026-10-12
CLAB-2026-774libgeos/geosRelateNG returns incorrect containment matrices for a valid tangent hole
libgeos/geos/issues/1558		Open2026-10-12
CLAB-2026-775google/s2geometryS2Cap::InteriorIntersects returns false for overlapping antipodal caps
google/s2geometry/issues/690		Open2026-10-12
CLAB-2026-777Xilinx/RapidWrightEDIF export collides renamed instance a-b with existing instance a_b
Xilinx/RapidWright/issues/1461		Open2026-10-12
CLAB-2026-785opencollab/arpack-ngSSAUPD/SSEUPD: surviving single-precision BE defect with RVEC=false
opencollab/arpack-ng/issues/499		Open2026-10-12
CLAB-2026-788ARM-software/astc-encoderHDR-to-8-bit output is rejected by the CLI and changes decoded API colors
ARM-software/astc-encoder/issues/675		Open2026-10-12
CLAB-2026-801coin-or/ClpDefault and direct dual report optimality for an unbounded LP with two free variables
coin-or/Clp/issues/341		Open2026-10-12
CLAB-2026-807kokkos/kokkos-kernelsNative tiled GEMM returns NaN for alpha=0 with finite inputs
kokkos/kokkos-kernels/issues/3335		Fixed2026-10-12
CLAB-2026-818qualcomm/aimetConstant prequantization rounds halfway weights after adding the zero point
qualcomm/aimet/issues/4124		Open2026-10-12
CLAB-2026-821intel/neural-compressorAsymmetric RTN INT8 recovery wraps centered codes
intel/neural-compressor/issues/2612		Open2026-10-12
CLAB-2026-829google/or-toolsFree MPS export rounds exact integer bounds to six significant digits
google/or-tools/issues/5481		Open2026-10-12
CLAB-2026-834AcademySoftwareFoundation/openexrOpenEXRCore B44/B44A writes planar UINT bytes when sizes tie
AcademySoftwareFoundation/openexr/issues/2728		Open2026-10-12
CLAB-2026-839enjoy-digital/litesataWRITE_DMA_EXT reports success for a DF-only terminal status
enjoy-digital/litesata/issues/40		Open2026-10-12
CLAB-2026-842scipopt/scipMPS UP-before-LO order discards an explicit integer upper bound
scipopt/scip/issues/231		Open2026-10-12
CLAB-2026-851huggingface/optimum-quantoDefault float8 quantization turns zero channels and calibration inputs into NaNs
huggingface/optimum-quanto/issues/437		Open2026-10-12
CLAB-2026-863apache/orc[C++] FLOAT statistics use values before binary32 narrowing
apache/orc/issues/2719		Open2026-10-12
CLAB-2026-867apache/arrow-nanoarrow[C] Half-float getter and appender return wrong values for nonzero subnormals
apache/arrow-nanoarrow/issues/961		Open2026-10-12
CLAB-2026-869google-ai-edge/LiteRTQuantized int8 DIV saturates small quotients to the int8 endpoints
google-ai-edge/LiteRT/issues/10745		Open2026-10-12
CLAB-2026-873PDAL/PDALwriters.las uses the wrong return limit for LAS 1.4 legacy formats and caches it across writers
PDAL/PDAL/issues/5145		Open2026-10-12
CLAB-2026-880xtensor-stack/xtensorAxis variance applies ddof before computing the per-lane sample count
xtensor-stack/xtensor/issues/2937		Open2026-10-12
CLAB-2026-887uxlfoundation/oneDPLDefault transform_reduce narrows a uint64_t initial value to the input type
uxlfoundation/oneDPL/issues/2851		Open2026-10-12
CLAB-2026-911mfem/mfemVTU cell attributes are selected by active scalars and array order
mfem/mfem/issues/5617		Open2026-10-12
CLAB-2026-912onnx/sklearn-onnxColumnTransformer loses column order when selecting all columns
onnx/sklearn-onnx/issues/1281		Fixed2026-10-12
CLAB-2026-922uxlfoundation/oneDALOnline moments return NaN dispersion after a one-row block
uxlfoundation/oneDAL/issues/3873		Open2026-10-12
CLAB-2026-927ginkgo-project/ginkgoSparsityCsr::transpose() resets a supplied shared value to 1
ginkgo-project/ginkgo/issues/2122		Open2026-10-12
CLAB-2026-929boostorg/histogramUnsigned integer axis maps below-range values to overflow
boostorg/histogram/issues/446		Open2026-10-12
CLAB-2026-941chipsalliance/sv-testsTwo unary operator simulation tests select elaboration because their type metadata lacks a colon
chipsalliance/sv-tests/issues/9218		Open2026-10-12
CLAB-2026-944protocolbuffers/protobufC++ ProtoJSON rounds quoted 64-bit decimal/exponent integers and rejects an in-range value
protocolbuffers/protobuf/issues/30634		Open2026-10-12
CLAB-2026-962FEniCS/basixReordered Legendre element returns interpolation coefficients in the reference order
FEniCS/basix/issues/1086		Open2026-10-12
CLAB-2026-157renode/renodeRISC-V Zvfh signed int16/binary16 conversions use 32-bit integer semantics at SEW=16
renode/renode/issues/1027		Open2026-10-13
CLAB-2026-171tenstorrent/whisperZvfh operations and FLI.H incorrectly require full scalar Zfh
tenstorrent/whisper/issues/125		Open2026-10-13
CLAB-2026-200riscv-mcu/e203_hbirdv2Machine external interrupts are selected after software and timer interrupts
riscv-mcu/e203_hbirdv2/issues/54		Open2026-10-13
CLAB-2026-893riscv-software-src/riscv-testsZfh format-conversion round trips pass when the conversions copy raw bits
riscv-software-src/riscv-tests/issues/670		Open2026-10-13
CLAB-2026-951szcompressor/SZ3CLI -A does not override the configured absolute bound unless -M is also supplied
szcompressor/SZ3/issues/187		Closed2026-10-13
CLAB-2026-977dmlc/xgboostExact interaction constraints skip a permitted feature when columns exceed rows
dmlc/xgboost/issues/12712		Open2026-10-13
CLAB-2026-1004capnproto/capnprotoJsonCodec loses a selected null Text union member when decoding its own output
capnproto/capnproto/issues/2856		Open2026-10-13
CLAB-2026-1013llnl/unitsDocumented unit-string addition returns 42 inches for 3 ft + 2 in
llnl/units/issues/475		Open2026-10-13
CLAB-2026-1027pvlib/pvlib-pythonDIRINT loses the nighttime zero and missing-dew-point fallback
pvlib/pvlib-python/issues/2900		Open2026-10-13
CLAB-2026-1032locationtech/jtsIsSimpleOp treats two empty MultiPoint members as duplicate coordinates
locationtech/jts/issues/1243		Open2026-10-13
CLAB-2026-1044cvxpy/cvxpyupper_tri returns a 1-D value for a declared column and evaluates matching-column addition incorrectly
cvxpy/cvxpy/issues/3558		Open2026-10-13
CLAB-2026-1075icl-utk-edu/blasppComplex CPU RowMajor her2 and mixed-type her reverse the imaginary off-diagonal update
icl-utk-edu/blaspp/issues/138		Open2026-10-13
CLAB-2026-1079Unidata/MetPySaturation calculations drop masked temperatures at two np.where sites
Unidata/MetPy/issues/4154		Open2026-10-13
CLAB-2026-1088sandialabs/seacasTRI4 shell top-face node list overwrites the third corner
sandialabs/seacas/issues/954		Open2026-10-13
CLAB-2026-1099SciTools/cf-unitsUnit.convert skips a 360_day origin offset and aliases identity conversions
SciTools/cf-units/issues/693		Open2026-10-13
CLAB-2026-1109AMReX-Codes/amrexCellConservativeLinear does not conserve volume integrals in 1D cylindrical geometry
AMReX-Codes/amrex/issues/6076		Open2026-10-13
CLAB-2026-1122scikit-hep/awkwardargmin and argmax renumber positions after missing lists
scikit-hep/awkward/issues/4426		Open2026-10-13
CLAB-2026-1154uber/h3greatCircleDistance returns NaN for valid antipodal coordinates
uber/h3/issues/1261		Open2026-10-13
CLAB-2026-1156healpy/healpyud_grade scales UNSEEN when upgrading with power=-2
healpy/healpy/issues/1134		Open2026-10-13
CLAB-2026-1160pymc-devs/pytensorBUG: Multi-axis sample variance subtracts ddof from each dimension
pymc-devs/pytensor/issues/2493		Open2026-10-13
CLAB-2026-1161kornia/kornia[Bug]: Horn essential-matrix decomposition returns NaN gradients at the identity pose
kornia/kornia/issues/5867		Fixed2026-10-13
CLAB-2026-1162keras-team/keraslinalg.solve misinterprets batched vector right-hand sides on NumPy and JAX
keras-team/keras/issues/23931		Open2026-10-13
CLAB-2026-1163cvxgrp/scsPower-cone solve returns SCS_SOLVED outside the requested 1e-8 feasibility tolerance
cvxgrp/scs/issues/491		Open2026-10-14
CLAB-2026-1165casadi/casadicopysign derivatives have the wrong sign for negative magnitude inputs
casadi/casadi/issues/4436		Open2026-10-14
CLAB-2026-1166google-deepmind/optaxclip_by_global_norm zeroes finite float16 updates
google-deepmind/optax/issues/1821		Open2026-10-14
CLAB-2026-1168pyro-ppl/numpyroLowRankMultivariateNormal.precision_matrix reverses the triangular-solve operands
pyro-ppl/numpyro/issues/2347		Open2026-10-14
CLAB-2026-1169dfki-ric/pytransform3dquaternions_from_matrices returns a non-unit quaternion near an x-axis half-turn
dfki-ric/pytransform3d/issues/399		Open2026-10-14
CLAB-2026-1172arviz-devs/arviz-statssummary pools MAD across chains when sample_dims=['draw']
arviz-devs/arviz-stats/issues/492		Open2026-10-14
CLAB-2026-1173ceres-solver/ceres-solverDENSE_SVD drops a nonzero covariance mode for a wide Jacobian with explicit null_space_rank
ceres-solver/ceres-solver/issues/1214		Open2026-10-14
CLAB-2026-1174CoolProp/CoolProp[ISSUE] IF97 molar H,S inputs overwrite the mass conversion
CoolProp/CoolProp/issues/3490		Open2026-10-14
CLAB-2026-1175JuliaStrings/utf8procUTF8PROC_CHARBOUND omits Hangul grapheme markers with COMPOSE or DECOMPOSE
JuliaStrings/utf8proc/issues/373		Open2026-10-14
CLAB-2026-1176JuliaStats/Distributions.jlTruncated discrete ccdf and logccdf return incorrect values at the lower atom
JuliaStats/Distributions.jl/issues/2095		Open2026-10-14
CLAB-2026-1177cornellius-gp/gpytorch[Bug] diag=True drops batch entries when batch size equals point count
cornellius-gp/gpytorch/issues/2801		Open2026-10-14
CLAB-2026-1178OpenMDAO/OpenMDAOBalanceComp constructor drops rhs_kwargs
OpenMDAO/OpenMDAO/issues/3828		Open2026-10-14
CLAB-2026-1179rai-opensource/spatialmath-pythontrlog2 loses the quadrant for valid SO(2) and SE(2) rotations
rai-opensource/spatialmath-python/issues/252		Open2026-10-14
CLAB-2026-1180spglib/spglibMagnetic datasets omit the Cartesian standardization rotation
spglib/spglib/issues/641		Open2026-10-14
CLAB-2026-1182dateutil/dateutilrrulestr drops TZID when VALUE follows it on DTSTART
dateutil/dateutil/issues/1611		Open2026-10-14
CLAB-2026-1185JuliaIntervals/IntervalArithmetic.jlUncertain decimal literal 0.03?3 loses its exact zero endpoint
JuliaIntervals/IntervalArithmetic.jl/issues/785		Open2026-10-14
CLAB-2026-1186picolibc/picolibcC23 narrowing division returns NaN for defined quotients
picolibc/picolibc/issues/1427		Open2026-10-14
CLAB-2026-1188quantumlib/CirqPauli arithmetic drops global_shift from integer-power Pauli gates
quantumlib/Cirq/issues/8450		Open2026-10-14
CLAB-2026-1254facebookresearch/faissclone_index changes IndexFlatIPPanorama vectors and search results
facebookresearch/faiss/issues/5735		Open2026-10-14
CLAB-2026-1263handsontable/hyperformula[Bug]: SUBSTITUTE expands dollar sequences in replacement text
handsontable/hyperformula/issues/1808		Open2026-10-14
CLAB-2026-1271bufbuild/protobuf-estoJson() emits the last name for aliased enum values
bufbuild/protobuf-es/issues/1569		Open2026-10-14
CLAB-2026-1284amazon-ion/ion-pythonion_equals ignores timestamp fractional precision above six digits
amazon-ion/ion-python/issues/458		Open2026-10-14
CLAB-2026-1430AOMediaCodec/libavifProposal: small checked-arithmetic helpers for size and offset computations
AOMediaCodec/libavif/issues/3415		Open2026-10-14