Leaderboard github.com
engineerchuan/wav2vec2.cpp
unreviewed-ai-risk
AI provenance
100%
explicit
- Confidence
- 70%
- Coverage
- 100%
Estimated share of AI-generated code. Explicit attribution is a lower bound.
Code quality
51/100
weak
- Confidence
- 90%
- Coverage
- 100%
Complexity, duplication, lint density, tests, CI, docs and review process.
Security
69/100
weak
- Confidence
- 85%
- Coverage
- 100%
Secrets, static analysis, vulnerable dependencies, CI and supply-chain posture.
AI provenance
100% explicit
Explicit AI attribution on 100% of first-party code added; an estimated 100% including velocity signals.
How the score is built
- Explicitly attributed share100%
- Unattributed share in AI-pace bursts0%
- Estimated AI-generated share100%
- AI tool configuration files0
- Work sessions at AI pace0
- Releases at AI pacenot measured
- Coding-style index (1.0+ means possible)0.18
Attributed AI tools
| Tool | Commits | Lines added | First seen | Last seen |
|---|---|---|---|---|
| Claude Code | 1 | 209,774 | 2025-12-26 | 2025-12-26 |
Lines added per month all AI-attributed share
- 2025-12209,774 · 100% AI
Findings in AI-attributed lines vs the rest
From git blame: 209,536 lines trace to AI-attributed commits, 0 to others. Too few lines in one group for a fair comparison. Reported, not scored.
| Findings per 1,000 lines | AI-attributed | Other | Ratio |
|---|---|---|---|
| Code quality | 0.06 | — | — |
| Security | 0.20 | — | — |
Evidence by analyzer
Explicit AI attributionok2 items
- Claude Code: 1 attributed commits, 209,774 lines added
- Claude Code: trailer 'Co-Authored-By: Claude'
- attributed_commit_share
- 100.0%
- attributed_commits
- 1
- attributed_loc
- 209,774
- attributed_loc_share
- 100.0%
- commits
- 1
- effective_loc
- 209,774
- tools_detected
- 1
AI tool configurationok
- artifacts
- 0
- tools_configured
- 0
- workflow_agents
- 0
Commit velocityok
- excluded_import_loc
- 209,774
- flagged_loc
- 0
- flagged_loc_share
- 0.0%
- flagged_units
- 0
- max_loc_per_hour
- 0
- measured_loc
- 0
- median_loc_per_hour
- 0
- pr_timed_units
- 0
- pre_ai_era_burst_loc
- 0
- pre_ai_era_bursts
- 0
- pre_ai_era_measured_loc
- 0
- work_units
- 0
Release velocityskipped
Coding styleok8 items
- step-by-step narration comment: // First, get batch results
- step-by-step narration comment: * Finally, an aclTensor object is created from the cached memory and returned.
- step-by-step narration comment: // Step 1.3: update theta_scale_tensor according to ext_factor or freq_scale.
- step-by-step narration comment: // Step 1.4: prepare select index if mrope
- step-by-step narration comment: // Step 5: multiply by attn_factor
- step-by-step narration comment: // Step 6: repeat
- step-by-step narration comment: // Step 1: Prepare trans tensors for F16 type conversion to F32 if needed
- step-by-step narration comment: // Step 2: Prepare head tensors for tail splitting if needed
- count:assistant_voice
- 0
- count:emoji_in_code
- 0
- count:narrated_steps
- 27
- count:placeholder
- 0
- files:assistant_voice
- 0
- files:emoji_in_code
- 0
- files:narrated_steps
- 7
- files:placeholder
- 0
- index
- 0.18
- lines_scanned
- 153,845
- per_kloc:assistant_voice
- 0
- per_kloc:emoji_in_code
- 0
- per_kloc:narrated_steps
- 0.18
- per_kloc:placeholder
- 0
Findings in AI-written linesok
- ai_lines
- 209,536
- blame_failures
- 0
- files_blamed
- 230
- other_lines
- 0
- quality:ai_findings
- 13
- quality:ai_per_kloc
- 0.06
- quality:other_findings
- 0
- security:ai_findings
- 41
- security:ai_per_kloc
- 0.20
- security:other_findings
- 0
Code quality
51 weak
Code quality is weak: test-to-source ratio 0.00, no CI test run detected; 41% of function code in complex or long functions; 23% duplicated lines. Weakest area: duplication.
How the score is built
- Complexity29
- Duplication26
- Lint and smells99
- Tests and CI45
- Documentation80
- Type safety40
- Review processnot measured
Evidence by analyzer
Complexity (lizard)ok10 items
- ggml_backend_vk_device_supports_op in ggml/src/ggml-vulkan/ggml-vulkan.cpp: complexity 307, 453 lines
- ggml_vk_op_get_pipeline in ggml/src/ggml-vulkan/ggml-vulkan.cpp: complexity 306, 544 lines
- detail::binary_reader::parse_msgpack_internal in examples/json.hpp: complexity 273, 340 lines
- detail::binary_reader::parse_cbor_internal in examples/json.hpp: complexity 239, 425 lines
- detail::lexer::scan_string in examples/json.hpp: complexity 212, 488 lines
- ggml_backend_sycl_device_supports_op in ggml/src/ggml-sycl/ggml-sycl.cpp: complexity 203, 278 lines
- ggml_vk_get_device in ggml/src/ggml-vulkan/ggml-vulkan.cpp: complexity 181, 547 lines
- ggml_cl_mul_mat in ggml/src/ggml-opencl/ggml-opencl.cpp: complexity 176, 788 lines
- ggml_metal_device_supports_op in ggml/src/ggml-metal/ggml-metal-device.m: complexity 167, 242 lines
- ggml_vk_load_shaders in ggml/src/ggml-vulkan/ggml-vulkan.cpp: complexity 167, 1073 lines
- complex_functions
- 268
- function_nloc
- 127,105
- functions
- 5,596
- long_functions
- 188
- many_param_functions
- 398
- mean_ccn
- 4.53
- p90_ccn
- 9
- risky_nloc_share
- 40.8%
Duplication (jscpd)ok10 items
- 6+ duplicated lines: examples/common-ggml.cpp:181 and ggml/src/ggml-sycl/ggml-sycl.cpp:3048
- 6+ duplicated lines: examples/common-ggml.cpp:203 and ggml/src/ggml-webgpu/ggml-webgpu.cpp:2380
- 6+ duplicated lines: examples/common.cpp:401 and examples/common.h:113
- 6+ duplicated lines: examples/common.cpp:420 and examples/common.cpp:535
- 6+ duplicated lines: examples/json.hpp:1 and examples/json.hpp:36
- 6+ duplicated lines: ggml/include/ggml-cann.h:2 and ggml/src/ggml-cann/acl_tensor.cpp:2
- 6+ duplicated lines: ggml/include/ggml-cann.h:63 and ggml/src/ggml-cann/ggml-cann.cpp:1497
- 6+ duplicated lines: ggml/include/ggml-opt.h:230 and ggml/src/ggml-opt.cpp:926
- 6+ duplicated lines: ggml/include/ggml.h:907 and ggml/include/ggml.h:916
- 6+ duplicated lines: ggml/include/ggml.h:1879 and ggml/src/ggml.c:4291
- compared_lines
- 150,528
- duplicated_lines
- 34,321
- duplicated_share
- 22.8%
- files
- 230
Lintok
- category:best-practice
- 8
- category:correctness
- 5
- findings
- 13
- scanned_lines
- 207,248
- weighted_findings
- 9.25
Tests and CIok2 items
- CI config: .github/workflows/build.yml
.github/workflows/build.yml - CI config: .github/workflows/eval.yml
.github/workflows/eval.yml
- check:ci_present
- yes
- check:ci_runs_lint
- no
- check:ci_runs_tests
- no
- check:has_tests
- yes
- check:test_ratio
- no
- check:tests_cochange
- yes
- ci_files
- 2
- cochange_commits
- 1
- cochange_share
- 100.0%
- source_lines
- 209,536
- test_lines
- 238
- test_ratio
- 0.00
Documentationok3 items
- README.md
README.md - LICENSE
LICENSE - CONTRIBUTING.md
CONTRIBUTING.md
- check:changelog
- no
- check:contributing
- yes
- check:license
- yes
- check:readme
- yes
- readme_lines
- 115
Type safetyok
- check:low_escape_density
- yes
- check:strict_config
- no
- escapes
- 0
- escapes_per_kloc
- 0
Security
69 weak
Security posture is weak: 2 high-severity SAST findings; 6 high-severity workflow issues. Weakest area: posture.
How the score is built
- Secrets100
- Static analysis (SAST)98
- Vulnerable dependenciesnot measured
- CI workflows23
- Security posture0
- OpenSSF Scorecardnot measured
Evidence by analyzer
OpenSSF Scorecardskipped
no published OpenSSF Scorecard result for this repo
Package registriesskipped
no registry dependencies declared
Secrets (gitleaks)ok
- secrets_current
- 0
- secrets_history
- 0
- secrets_low_confidence
- 0
- secrets_nonprod
- 0
Static analysis (opengrep)ok
- findings
- 41
- findings_high
- 2
- findings_low
- 0
- findings_medium
- 39
- parse_errors
- 40
- scanned_lines
- 207,248
- weighted_findings
- 11.25
Dependencies (osv-scanner)skipped
CI workflows (zizmor)ok
- findings
- 15
- findings_high
- 6
- findings_low
- 0
- findings_medium
- 9
- penalty
- 88.80
- workflows
- 2
Security postureok
- check:dependency_updates
- no
- check:lockfiles
- yes
- check:security_policy
- no
- check:signed_commits
- no
- signed_commit_share
- 0.0%
- unpinned_manifests
- 0
Release downloadsok
- binary_assets
- 0
- signals
- 0
- suspicious
- 0
Findings
8 high56 medium5 low
69 findings
| Severity | Rule | Location | Message | Tool |
|---|---|---|---|---|
| high | zizmor/unpinned-uses |
.github/workflows/build.yml:15 | unpinned action reference [confidence high] | CI workflows (zizmor) |
| high | zizmor/unpinned-uses |
.github/workflows/build.yml:40 | unpinned action reference [confidence high] | CI workflows (zizmor) |
| high | zizmor/unpinned-uses |
.github/workflows/build.yml:60 | unpinned action reference [confidence high] | CI workflows (zizmor) |
| high | zizmor/unpinned-uses |
.github/workflows/eval.yml:21 | unpinned action reference [confidence high] | CI workflows (zizmor) |
| high | zizmor/unpinned-uses |
.github/workflows/eval.yml:26 | unpinned action reference [confidence high] | CI workflows (zizmor) |
| high | zizmor/unpinned-uses |
.github/workflows/eval.yml:86 | unpinned action reference [confidence high] | CI workflows (zizmor) |
| high | opengrep/python.lang.security.audit.dangerous-subprocess-use-audit |
scripts/eval_vs_pytorch.py:109 | Detected subprocess function 'run' without a static string. If this data can be controlled by a malicious actor, it may be an instance of command injection. Audit the use of this call to ensure it is not controllable by an external... | Static analysis (opengrep) |
| high | opengrep/python.lang.security.audit.dangerous-subprocess-use-audit |
scripts/eval_vs_pytorch.py:118 | Detected subprocess function 'run' without a static string. If this data can be controlled by a malicious actor, it may be an instance of command injection. Audit the use of this call to ensure it is not controllable by an external... | Static analysis (opengrep) |
| medium | opengrep/python.lang.best-practice.unspecified-open-encoding |
ggml/src/ggml-cuda/template-instances/generate_cu_files.py:65 | Missing 'encoding' parameter. 'open()' uses device locale encodings by default, corrupting files with special characters. Specify the encoding to ensure cross-platform support when opening files in text mode (e.g. encoding="utf-8"). | Lint |
| medium | opengrep/python.lang.best-practice.unspecified-open-encoding |
ggml/src/ggml-cuda/template-instances/generate_cu_files.py:70 | Missing 'encoding' parameter. 'open()' uses device locale encodings by default, corrupting files with special characters. Specify the encoding to ensure cross-platform support when opening files in text mode (e.g. encoding="utf-8"). | Lint |
| medium | opengrep/python.lang.best-practice.unspecified-open-encoding |
ggml/src/ggml-cuda/template-instances/generate_cu_files.py:78 | Missing 'encoding' parameter. 'open()' uses device locale encodings by default, corrupting files with special characters. Specify the encoding to ensure cross-platform support when opening files in text mode (e.g. encoding="utf-8"). | Lint |
| medium | opengrep/python.lang.best-practice.unspecified-open-encoding |
ggml/src/ggml-cuda/template-instances/generate_cu_files.py:94 | Missing 'encoding' parameter. 'open()' uses device locale encodings by default, corrupting files with special characters. Specify the encoding to ensure cross-platform support when opening files in text mode (e.g. encoding="utf-8"). | Lint |
| medium | opengrep/python.lang.best-practice.unspecified-open-encoding |
ggml/src/ggml-cuda/template-instances/generate_cu_files.py:98 | Missing 'encoding' parameter. 'open()' uses device locale encodings by default, corrupting files with special characters. Specify the encoding to ensure cross-platform support when opening files in text mode (e.g. encoding="utf-8"). | Lint |
| medium | opengrep/python.lang.best-practice.unspecified-open-encoding |
ggml/src/ggml-opencl/kernels/embed_kernel.py:15 | Missing 'encoding' parameter. 'open()' uses device locale encodings by default, corrupting files with special characters. Specify the encoding to ensure cross-platform support when opening files in text mode (e.g. encoding="utf-8"). | Lint |
| medium | opengrep/python.lang.best-practice.unspecified-open-encoding |
ggml/src/ggml-opencl/kernels/embed_kernel.py:16 | Missing 'encoding' parameter. 'open()' uses device locale encodings by default, corrupting files with special characters. Specify the encoding to ensure cross-platform support when opening files in text mode (e.g. encoding="utf-8"). | Lint |
| medium | opengrep/python.lang.best-practice.unspecified-open-encoding |
scripts/eval_vs_pytorch.py:374 | Missing 'encoding' parameter. 'open()' uses device locale encodings by default, corrupting files with special characters. Specify the encoding to ensure cross-platform support when opening files in text mode (e.g. encoding="utf-8"). | Lint |
| medium | zizmor/excessive-permissions |
.github/workflows/build.yml:1 | overly broad permissions [confidence medium] | CI workflows (zizmor) |
| medium | zizmor/excessive-permissions |
.github/workflows/build.yml:11 | overly broad permissions [confidence medium] | CI workflows (zizmor) |
| medium | zizmor/artipacked |
.github/workflows/build.yml:15 | credential persistence through GitHub Actions artifacts [confidence low] | CI workflows (zizmor) |
| medium | zizmor/excessive-permissions |
.github/workflows/build.yml:36 | overly broad permissions [confidence medium] | CI workflows (zizmor) |
| medium | zizmor/artipacked |
.github/workflows/build.yml:40 | credential persistence through GitHub Actions artifacts [confidence low] | CI workflows (zizmor) |
| medium | zizmor/excessive-permissions |
.github/workflows/build.yml:56 | overly broad permissions [confidence medium] | CI workflows (zizmor) |
| medium | zizmor/artipacked |
.github/workflows/build.yml:60 | credential persistence through GitHub Actions artifacts [confidence low] | CI workflows (zizmor) |
| medium | zizmor/excessive-permissions |
.github/workflows/eval.yml:17 | overly broad permissions [confidence medium] | CI workflows (zizmor) |
| medium | zizmor/artipacked |
.github/workflows/eval.yml:21 | credential persistence through GitHub Actions artifacts [confidence low] | CI workflows (zizmor) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-alloc.c:724 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.function-use-after-free |
ggml/src/ggml-alloc.c:857 | Variable 'galloc->node_allocs' was passed to a function after being freed. This can lead to undefined behavior. [confidence low] | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.function-use-after-free |
ggml/src/ggml-alloc.c:889 | Variable 'galloc->leaf_allocs' was passed to a function after being freed. This can lead to undefined behavior. [confidence low] | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.use-after-free |
ggml/src/ggml-alloc.c:889 | Variable 'galloc->leaf_allocs' was used after being freed. This can lead to undefined behavior. [confidence low] | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.function-use-after-free |
ggml/src/ggml-alloc.c:890 | Variable 'galloc->leaf_allocs' was passed to a function after being freed. This can lead to undefined behavior. [confidence low] | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-cpu/ggml-cpu.c:1590 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-cpu/ggml-cpu.c:2626 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-cpu/ggml-cpu.c:2733 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-cpu/ggml-cpu.c:3132 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-cpu/quants.c:499 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-cpu/quants.c:509 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-cpu/quants.c:576 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-cpu/quants.c:582 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-cpu/quants.c:651 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-cpu/quants.c:658 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-cpu/quants.c:722 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-cpu/quants.c:729 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-hexagon/htp/htp-dma.c:30 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-hexagon/htp/htp-dma.c:35 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-hexagon/htp/htp-dma.c:38 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-hexagon/htp/main.c:49 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-hexagon/htp/main.c:60 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-hexagon/htp/main.c:80 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-hexagon/htp/main.c:91 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-hexagon/htp/matmul-ops.c:1694 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-hexagon/htp/matmul-ops.c:2201 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-hexagon/htp/rope-ops.c:128 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml-hexagon/htp/softmax-ops.c:76 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-strcat-fn |
ggml/src/ggml-hexagon/htp/worker-pool.c:97 | Finding triggers whenever there is a strcat or strncat used. This is an issue because strcat or strncat can lead to buffer overflow vulns. Fix this by using strcat_s instead. [confidence low] | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml.c:4098 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml.c:6108 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml.c:6807 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml.c:6808 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml.c:6966 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml.c:6967 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml.c:7020 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml.c:7021 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml.c:7055 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| medium | opengrep/c.lang.security.insecure-use-memset |
ggml/src/ggml.c:7585 | When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While `memset()` is commonly used for this purpose, it can leave sensitive information... | Static analysis (opengrep) |
| low | opengrep/bash.lang.correctness.unquoted-variable-expansion-in-command |
build-wav2vec2-xcframework.sh:125 | Variable expansions must be double-quoted so as to prevent being split into multiple pieces according to whitespace or whichever separator is specified by the IFS variable. If you really wish to split the variable's contents, you may... | Lint |
| low | opengrep/bash.lang.correctness.unquoted-command-substitution-in-command |
build-wav2vec2-xcframework.sh:127 | The result of command substitution $(...) or `...`, if unquoted, is split on whitespace or other separators specified by the IFS variable. You should surround it with double quotes to avoid splitting the result. | Lint |
| low | opengrep/bash.lang.correctness.unquoted-variable-expansion-in-command |
build-wav2vec2-xcframework.sh:127 | Variable expansions must be double-quoted so as to prevent being split into multiple pieces according to whitespace or whichever separator is specified by the IFS variable. If you really wish to split the variable's contents, you may... | Lint |
| low | opengrep/bash.lang.correctness.unquoted-variable-expansion-in-command |
build-wav2vec2-xcframework.sh:49 | Variable expansions must be double-quoted so as to prevent being split into multiple pieces according to whitespace or whichever separator is specified by the IFS variable. If you really wish to split the variable's contents, you may... | Lint |
| low | opengrep/bash.lang.correctness.unquoted-variable-expansion-in-command |
build-wav2vec2-xcframework.sh:63 | Variable expansions must be double-quoted so as to prevent being split into multiple pieces according to whitespace or whichever separator is specified by the IFS variable. If you really wish to split the variable's contents, you may... | Lint |
Tools and coverage
Every analyzer that ran on this commit. A failed or skipped tool is a coverage gap: its share of the score is left out and the coverage figure drops.
| Analyzer | Pillar | Status | Version | Findings | Note |
|---|---|---|---|---|---|
| AI tool configuration | AI provenance | ok | signatures-v1 | 0 | |
| Explicit AI attribution | AI provenance | ok | signatures-v1 | 0 | |
| Release velocity | AI provenance | skipped | — | 0 | |
| Coding style | AI provenance | ok | — | 0 | |
| Commit velocity | AI provenance | ok | prov-4 | 0 | |
| Findings in AI-written lines | AI provenance | ok | git-blame | 0 | |
| Complexity (lizard) | Code quality | ok | lizard 1.25.1 | 0 | |
| Documentation | Code quality | ok | — | 0 | |
| Duplication (jscpd) | Code quality | ok | repotrust-dup-w6 | 0 | |
| Lint | Code quality | ok | opengrep 1.33.1; rules f1d2b562b414 | 13 | |
| Tests and CI | Code quality | ok | — | 0 | |
| Type safety | Code quality | ok | — | 0 | |
| CI workflows (zizmor) | Security | ok | zizmor 1.32.0 | 15 | |
| Dependencies (osv-scanner) | Security | skipped | — | 0 | |
| Release downloads | Security | ok | — | 0 | |
| Security posture | Security | ok | — | 0 | |
| Package registries | Security | skipped | — | 0 | no registry dependencies declared |
| Static analysis (opengrep) | Security | ok | opengrep 1.33.1; rules f1d2b562b414 | 41 | |
| OpenSSF Scorecard | Security | skipped | — | 0 | no published OpenSSF Scorecard result for this repo |
| Secrets (gitleaks) | Security | ok | gitleaks 8.30.1 | 0 |
Score history
Scores from different scoring versions aren't directly comparable.
| Scored | Commit | Scoring | AI provenance | Quality | Security |
|---|---|---|---|---|---|
| 2026-11-14 | db8d0f108a89 | prov-5+qual-2+sec-2 | 100% | 51 | 69 |
| 2026-11-14 | db8d0f108a89 | prov-4+qual-2+sec-2 | 100% | 51 | 69 |
Full text report
- Repository: https://github.com/engineerchuan/wav2vec2.cpp
- Commit:
db8d0f108a89fe030f4603e7d8cf6006d1978968 - Generated: 2026-11-14T14:52:04+00:00
- Scorer 0.1.0, scoring prov-5+qual-2+sec-2, report schema v2
Scores
Unreviewed AI code risk: a large share of this code is explicitly AI-attributed, and its tests or review process are weak while quality or security scores are low.
| Pillar | Score | Signal | Confidence | Coverage |
|---|---|---|---|---|
| AI provenance (estimated AI-generated share) | 100 | explicit | 70% | 100% |
| Code quality | 51 | weak | 90% | 100% |
| Security | 69 | weak | 85% | 100% |
AI provenance
Explicit AI attribution on 100% of first-party code added; an estimated 100% including velocity signals.
- Explicitly attributed share (lower bound): 100.0%
- Unattributed share in AI-consistent velocity bursts: 0.0%
- Estimated AI-generated share: 100.0%
| AI tool | Attributed commits | Lines added |
|---|---|---|
| Claude Code | 1 | 209,774 |
Current source lines last changed by AI-attributed commits: 209,536 of 209,536.
Too few lines in one group to compare finding densities.
- Stylometry index: 0.18 per 1,000 lines (narrated steps 27, placeholders 0, assistant phrasing 0, emoji 0). Reported as a signal, not scored.
| Month | Lines added | Attributed to AI |
|---|---|---|
| 2025-12 | 209,774 | 100% |
Evidence
- Claude Code: trailer 'Co-Authored-By: Claude' (db8d0f108a89)
- step-by-step narration comment: // First, get batch results (examples/wav)
- step-by-step narration comment: * Finally, an aclTensor object is created from the cached memory and returned. (ggml/src/ggm)
- step-by-step narration comment: // Step 1.3: update theta_scale_tensor according to ext_factor or freq_scale. (ggml/src/ggm)
- step-by-step narration comment: // Step 1.4: prepare select index if mrope (ggml/src/ggm)
- step-by-step narration comment: // Step 5: multiply by attn_factor (ggml/src/ggm)
- step-by-step narration comment: // Step 6: repeat (ggml/src/ggm)
- step-by-step narration comment: // Step 1: Prepare trans tensors for F16 type conversion to F32 if needed (ggml/src/ggm)
- step-by-step narration comment: // Step 2: Prepare head tensors for tail splitting if needed (ggml/src/ggm)
AI provenance describes how code was produced; it is not a quality judgment. Unless attribution is explicit, these are signals consistent with AI generation, not proof. Squash/rebase workflows, imported code, and rewritten history reduce accuracy.
Code quality
Code quality is weak: test-to-source ratio 0.00, no CI test run detected; 41% of function code in complex or long functions; 23% duplicated lines. Weakest area: duplication.
| Area | Score |
|---|---|
| Complexity (lizard) | 29 |
| Duplication | 26 |
| Code smells (Opengrep) | 99 |
| Tests and CI | 45 |
| Documentation | 80 |
| Type safety | 40 |
| Process (Scorecard review/maintenance) | n/a |
- Functions: 5,596; mean complexity 4.5, 90th percentile 9; 268 complex and 188 long functions
- Duplicated lines: 22.8% of 150,528 compared
- Tests: 238 test lines for 209,536 source lines (ratio 0.00); CI test commands: none detected; 100% of recent source commits also change tests
- Docs missing or thin: changelog
Most complex functions:
- ggml_backend_vk_device_supports_op in ggml/src/ggml-vulkan/ggml-vulkan.cpp: complexity 307, 453 lines
- ggml_vk_op_get_pipeline in ggml/src/ggml-vulkan/ggml-vulkan.cpp: complexity 306, 544 lines
- detail::binary_reader::parse_msgpack_internal in examples/json.hpp: complexity 273, 340 lines
- detail::binary_reader::parse_cbor_internal in examples/json.hpp: complexity 239, 425 lines
- detail::lexer::scan_string in examples/json.hpp: complexity 212, 488 lines
Security
Security posture is weak: 2 high-severity SAST findings; 6 high-severity workflow issues. Weakest area: posture.
| Area | Score |
|---|---|
| Secrets (gitleaks, tree + history) | 100 |
| Static analysis (Opengrep) | 98 |
| Vulnerable dependencies (OSV) | n/a |
| CI workflows (zizmor) | 23 |
| Posture (policy, updates, signing, lockfiles) | 0 |
| OpenSSF Scorecard | n/a |
- Posture checks passed: lockfiles; missing: security policy, dependency updates, signed commits
Top findings
| Severity | Finding | Location |
|---|---|---|
| high | opengrep/python.lang.security.audit.dangerous-subprocess-use-audit: Detected subprocess function 'run' without a static string. If this data can be controlled by a malicious actor, it may be an instance of command injection. Audit the use of this call to ensure it is not controllable by an external resource. You may consider using 'shlex.escape()'. [confidence low] |
scripts/eval_vs_pytorch.py:109 |
| high | opengrep/python.lang.security.audit.dangerous-subprocess-use-audit: Detected subprocess function 'run' without a static string. If this data can be controlled by a malicious actor, it may be an instance of command injection. Audit the use of this call to ensure it is not controllable by an external resource. You may consider using 'shlex.escape()'. [confidence low] |
scripts/eval_vs_pytorch.py:118 |
| high | zizmor/unpinned-uses: unpinned action reference [confidence high] |
.github/workflows/build.yml:15 |
| high | zizmor/unpinned-uses: unpinned action reference [confidence high] |
.github/workflows/build.yml:40 |
| high | zizmor/unpinned-uses: unpinned action reference [confidence high] |
.github/workflows/build.yml:60 |
| high | zizmor/unpinned-uses: unpinned action reference [confidence high] |
.github/workflows/eval.yml:21 |
| high | zizmor/unpinned-uses: unpinned action reference [confidence high] |
.github/workflows/eval.yml:26 |
| high | zizmor/unpinned-uses: unpinned action reference [confidence high] |
.github/workflows/eval.yml:86 |
| medium | opengrep/c.lang.security.insecure-use-memset: When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While memset() is commonly used for this purpose, it can leave sensitive information behind due to compiler optimizations or other factors. To avoi [confidence low] |
ggml/src/ggml-alloc.c:724 |
| medium | opengrep/c.lang.security.function-use-after-free: Variable 'galloc->node_allocs' was passed to a function after being freed. This can lead to undefined behavior. [confidence low] |
ggml/src/ggml-alloc.c:857 |
| medium | opengrep/c.lang.security.function-use-after-free: Variable 'galloc->leaf_allocs' was passed to a function after being freed. This can lead to undefined behavior. [confidence low] |
ggml/src/ggml-alloc.c:889 |
| medium | opengrep/c.lang.security.use-after-free: Variable 'galloc->leaf_allocs' was used after being freed. This can lead to undefined behavior. [confidence low] |
ggml/src/ggml-alloc.c:889 |
| medium | opengrep/c.lang.security.function-use-after-free: Variable 'galloc->leaf_allocs' was passed to a function after being freed. This can lead to undefined behavior. [confidence low] |
ggml/src/ggml-alloc.c:890 |
| medium | opengrep/c.lang.security.insecure-use-memset: When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While memset() is commonly used for this purpose, it can leave sensitive information behind due to compiler optimizations or other factors. To avoi [confidence low] |
ggml/src/ggml-cpu/ggml-cpu.c:1590 |
| medium | opengrep/c.lang.security.insecure-use-memset: When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While memset() is commonly used for this purpose, it can leave sensitive information behind due to compiler optimizations or other factors. To avoi [confidence low] |
ggml/src/ggml-cpu/ggml-cpu.c:2626 |
| medium | opengrep/c.lang.security.insecure-use-memset: When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While memset() is commonly used for this purpose, it can leave sensitive information behind due to compiler optimizations or other factors. To avoi [confidence low] |
ggml/src/ggml-cpu/ggml-cpu.c:2733 |
| medium | opengrep/c.lang.security.insecure-use-memset: When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While memset() is commonly used for this purpose, it can leave sensitive information behind due to compiler optimizations or other factors. To avoi [confidence low] |
ggml/src/ggml-cpu/ggml-cpu.c:3132 |
| medium | opengrep/c.lang.security.insecure-use-memset: When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While memset() is commonly used for this purpose, it can leave sensitive information behind due to compiler optimizations or other factors. To avoi [confidence low] |
ggml/src/ggml-cpu/quants.c:499 |
| medium | opengrep/c.lang.security.insecure-use-memset: When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While memset() is commonly used for this purpose, it can leave sensitive information behind due to compiler optimizations or other factors. To avoi [confidence low] |
ggml/src/ggml-cpu/quants.c:509 |
| medium | opengrep/c.lang.security.insecure-use-memset: When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While memset() is commonly used for this purpose, it can leave sensitive information behind due to compiler optimizations or other factors. To avoi [confidence low] |
ggml/src/ggml-cpu/quants.c:576 |
| medium | opengrep/c.lang.security.insecure-use-memset: When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While memset() is commonly used for this purpose, it can leave sensitive information behind due to compiler optimizations or other factors. To avoi [confidence low] |
ggml/src/ggml-cpu/quants.c:582 |
| medium | opengrep/c.lang.security.insecure-use-memset: When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While memset() is commonly used for this purpose, it can leave sensitive information behind due to compiler optimizations or other factors. To avoi [confidence low] |
ggml/src/ggml-cpu/quants.c:651 |
| medium | opengrep/c.lang.security.insecure-use-memset: When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While memset() is commonly used for this purpose, it can leave sensitive information behind due to compiler optimizations or other factors. To avoi [confidence low] |
ggml/src/ggml-cpu/quants.c:658 |
| medium | opengrep/c.lang.security.insecure-use-memset: When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While memset() is commonly used for this purpose, it can leave sensitive information behind due to compiler optimizations or other factors. To avoi [confidence low] |
ggml/src/ggml-cpu/quants.c:722 |
| medium | opengrep/c.lang.security.insecure-use-memset: When handling sensitive information in a buffer, it's important to ensure that the data is securely erased before the buffer is deleted or reused. While memset() is commonly used for this purpose, it can leave sensitive information behind due to compiler optimizations or other factors. To avoi [confidence low] |
ggml/src/ggml-cpu/quants.c:729 |
Static signals only: the scorer never runs the project's code. Secret findings are pattern matches that were not verified, and secret values are never stored.
Repository profile
- Source lines: 209,536; test lines: 238
| Language | Lines |
|---|---|
| C++ | 144,643 |
| C | 61,397 |
| Objective-C | 2,288 |
| Python | 1,172 |
| Shell | 274 |