CVE-2026-72852

HIGHPre-NVD 7.87.8—
EchelonGraph scoreMEDIUM confidence

Score 7.8 from GitHub Security Advisory (severity: HIGH) published 2026-08-20. A secondary CVSS source baseline 7.8; sources differ by 0.0.

Triggered by: GitHub Security Advisory CVSS
Sources: epss, ghsa, secondary
Trending — 3 sources updated this weekElevated
7.8EG
EchelonGraph verdictPlan mitigationSerious severity, but no confirmed exploitation yet.
  • High severity, but no confirmed exploitation yet
CISA-KEV: Not listedEPSS PROB: 0.2%CVSS: 7.8Exploit: Elevated riskExposed services: Not assessed

No fix is confirmed yet — apply a workaround or compensating control (WAF / firewall / segmentation) and watch for the fix.

hank-ai/darknet sizes a convolutional layer's weight and output heap buffers by multiplying configuration fields taken from a .cfg file in unchecked 32-bit int arithmetic. In src-lib/convolutional_layer.cpp, l.nweights is computed as (c / groups) * n * size * size and l.outputs as l.out_h * l.out_w * l.out_c, and both feed xcalloc directly. A .cfg whose true dimension product exceeds INT_MAX wraps to a small or zero value, so the allocation is undersized; for example width and height of 256 with filters of 65536 gives 2^32, which wraps to 0. forward_convolutional_layer then re-derives the GEMM dimensions with a different operand order, computing k as l.size*l.size*l.c / l.groups where the allocation divided before multiplying, and reads and writes through the undersized buffer. Loading the crafted .cfg for inference or training is sufficient and no valid .weights file is required. The reported proof of concept observed a heap buffer overflow read in gemm_nn_fast under AddressSanitizer and glibc allocator metadata corruption in a release build of the same input, indicating an out-of-bounds write.

CVSS v3
7.8
EG Score
7.8HIGHmedium confidence
EG Risk
51
EG Risk 51/100CISA SSVC

EG Risk is EchelonGraph's 0–100 priority score: it fuses intrinsic severity with real-world exploitation and automatability so you can rank equal-severity CVEs and fix the most dangerous first. Higher = act sooner. Distinct from the 0–10 EG Score (severity).

How it’s computed
Severity78% × 45%
Exploitation40% × 40%
Automatability0% × 15%
CISA SSVC: Track at low, Track* at medium and Attend at high mission impact.
Action: No fix is confirmed yet. Restrict network exposure of the affected system or apply the vendor's mitigation within your standard update timelines at low or medium mission impact and sooner than that at high, and watch the vendor's advisory for the fix.
EPSS PROB
0.2%
EPSS %ILE
10th
KEV
Not listed

CISA SSVCTrack at low, Track* at medium and Attend at high mission impact.

No fix is confirmed yet. Restrict network exposure of the affected system or apply the vendor's mitigation within your standard update timelines at low or medium mission impact and sooner than that at high, and watch the vendor's advisory for the fix.

Exploitation public PoC (CISA Vulnrichment) · Automatable no (CISA Vulnrichment) · Technical impact total (CISA Vulnrichment). Mission impact is CISA's Mission & Well-being decision point, and only you can judge it: high means the affected system is essential to your organisation's mission, or its compromise could cause irreversible harm to people. CISA's decision table

Published

August 20, 2026

Last Modified

September 24, 2026

Advisory Details (6)

Auto-updated Aug 20, 2026
🔬 Proof of concept available. No patch confirmed yet.
generic🟡 PoC Available

darknet Integer Overflow in Convolutional Layer Buffer Sizing Leads to Heap Buffer Overflow | Advisories | VulnCheck

https://www.vulncheck.com/advisories/darknet-integer-overflow-in-convolutional-layer-buffer-sizing-leads-to-heap-buffer-overflow
generic

GitHub - hank-ai/darknet: Darknet/YOLO object detection framework · GitHub

https://github.com/hank-ai/darknet
generic

darknet/src-lib/convolutional_layer.cpp at v6.0 · hank-ai/darknet · GitHub

https://github.com/hank-ai/darknet/blob/v6.0/src-lib/convolutional_layer.cpp#L1457
generic

darknet/src-lib/convolutional_layer.cpp at v6.0 · hank-ai/darknet · GitHub

https://github.com/hank-ai/darknet/blob/v6.0/src-lib/convolutional_layer.cpp#L811
generic

darknet/src-lib/convolutional_layer.cpp at v6.0 · hank-ai/darknet · GitHub

https://github.com/hank-ai/darknet/blob/v6.0/src-lib/convolutional_layer.cpp#L764
generic

Integer overflow in convolutional weight/output buffer sizing causes heap corruption · Issue #148 · hank-ai/darknet · GitHub

https://github.com/hank-ai/darknet/issues/148

Vendor Advisories for CVE-2026-72852(1)

These vendors published their own advisory mentioning this CVE — often with vendor-specific remediation steps + affected product lists not in NVD.

Weakness Classification(2)

MITRE Common Weakness Enumeration — the root-cause categories this CVE belongs to.

Data Freshness Timeline

(refreshed 21× in last 7d / 100× in last 30d)

Each row is a source pipeline that fetched or updated this CVE on that date, with what changed. For example, "NVD update" means NVD published or revised its analysis for this CVE; "MITRE cvelistV5" means we ingested or refreshed it from the CNA feed. Most recent first.

Showing the most recent 100 of 165 total refreshes for this CVE.

  1. 2026-10-05 09:32 UTCEG score recompute
  2. 2026-10-05 09:32 UTCGHSA enrichment
  3. 2026-10-04 23:22 UTCEPSS rescore
  4. 2026-10-04 20:52 UTCGHSA enrichment
  5. 2026-10-04 05:26 UTCGHSA enrichment
  6. 2026-10-03 16:47 UTCEG score recompute
  7. 2026-10-03 16:47 UTCGHSA enrichment
  8. 2026-10-03 14:26 UTCEPSS rescore
  9. 2026-10-03 01:36 UTCEG score recompute
  10. 2026-10-03 01:36 UTCGHSA enrichment
  11. 2026-10-02 12:55 UTCGHSA enrichment
  12. 2026-10-01 22:44 UTCEG score recompute
  13. 2026-10-01 22:44 UTCGHSA enrichment
  14. 2026-10-01 19:50 UTCEPSS rescore
  15. 2026-10-01 09:33 UTCGHSA enrichment
  16. 2026-09-30 20:52 UTCEG score recompute
  17. 2026-09-30 20:52 UTCGHSA enrichment
  18. 2026-09-30 08:12 UTCEG score recompute
  19. 2026-09-30 08:12 UTCGHSA enrichment
  20. 2026-09-28 21:38 UTCEG score recompute
  21. 2026-09-28 21:38 UTCGHSA enrichment
  22. 2026-09-28 13:52 UTCEPSS rescore
  23. 2026-09-28 04:39 UTCGHSA enrichment
  24. 2026-09-27 15:56 UTCEG score recompute
  25. 2026-09-27 15:55 UTCGHSA enrichment
Show 75 more
  1. 2026-09-27 13:48 UTCEPSS rescore
  2. 2026-09-27 02:59 UTCEG score recompute
  3. 2026-09-27 02:59 UTCGHSA enrichment
  4. 2026-09-26 15:59 UTCEPSS rescore
  5. 2026-09-25 10:01 UTCGHSA enrichment
  6. 2026-09-24 21:21 UTCEG score recompute
  7. 2026-09-24 21:21 UTCGHSA enrichment
  8. 2026-09-24 14:58 UTCEG score recompute
  9. 2026-09-24 14:58 UTCGHSA enrichment
  10. 2026-09-24 10:57 UTCGHSA enrichment
  11. 2026-09-23 22:17 UTCEG score recompute
  12. 2026-09-23 22:17 UTCGHSA enrichment
  13. 2026-09-23 17:54 UTCEPSS rescore
  14. 2026-09-23 09:37 UTCGHSA enrichment
  15. 2026-09-22 20:57 UTCEG score recompute
  16. 2026-09-22 20:57 UTCGHSA enrichment
  17. 2026-09-22 08:16 UTCEG score recompute
  18. 2026-09-22 08:16 UTCGHSA enrichment
  19. 2026-09-21 21:09 UTCEPSS rescore
  20. 2026-09-21 19:36 UTCGHSA enrichment
  21. 2026-09-21 06:56 UTCEG score recompute
  22. 2026-09-21 06:55 UTCGHSA enrichment
  23. 2026-09-20 20:16 UTCEPSS rescore
  24. 2026-09-20 18:14 UTCGHSA enrichment
  25. 2026-09-20 05:21 UTCEG score recompute
  26. 2026-09-20 05:21 UTCGHSA enrichment
  27. 2026-09-19 16:40 UTCEG score recompute
  28. 2026-09-19 16:40 UTCGHSA enrichment
  29. 2026-09-19 03:59 UTCEG score recompute
  30. 2026-09-19 03:59 UTCGHSA enrichment
  31. 2026-09-18 19:28 UTCEPSS rescore
  32. 2026-09-18 15:18 UTCGHSA enrichment
  33. 2026-09-18 02:38 UTCEG score recompute
  34. 2026-09-18 02:38 UTCGHSA enrichment
  35. 2026-09-17 13:57 UTCGHSA enrichment
  36. 2026-09-17 01:17 UTCEG score recompute
  37. 2026-09-17 01:17 UTCGHSA enrichment
  38. 2026-09-16 14:08 UTCEPSS rescore
  39. 2026-09-16 12:36 UTCEG score recompute
  40. 2026-09-16 12:36 UTCGHSA enrichment
  41. 2026-09-16 05:15 UTCEPSS rescore
  42. 2026-09-15 23:56 UTCGHSA enrichment
  43. 2026-09-15 11:15 UTCEG score recompute
  44. 2026-09-15 11:15 UTCGHSA enrichment
  45. 2026-09-15 03:12 UTCEPSS rescore
  46. 2026-09-14 21:58 UTCGHSA enrichment
  47. 2026-09-14 09:17 UTCGHSA enrichment
  48. 2026-09-13 20:36 UTCEG score recompute
  49. 2026-09-13 20:36 UTCGHSA enrichment
  50. 2026-09-13 07:55 UTCGHSA enrichment
  51. 2026-09-12 19:15 UTCEG score recompute
  52. 2026-09-12 19:15 UTCGHSA enrichment
  53. 2026-09-12 15:02 UTCEPSS rescore
  54. 2026-09-12 06:35 UTCGHSA enrichment
  55. 2026-09-11 17:55 UTCEG score recompute
  56. 2026-09-11 17:55 UTCGHSA enrichment
  57. 2026-09-11 14:53 UTCEPSS rescore
  58. 2026-09-11 05:14 UTCGHSA enrichment
  59. 2026-09-10 16:34 UTCEG score recompute
  60. 2026-09-10 16:34 UTCGHSA enrichment
  61. 2026-09-10 09:35 UTCEPSS rescore
  62. 2026-09-10 03:19 UTCGHSA enrichment
  63. 2026-09-09 14:38 UTCGHSA enrichment
  64. 2026-09-09 01:55 UTCEG score recompute
  65. 2026-09-09 01:55 UTCGHSA enrichment
  66. 2026-09-08 22:01 UTCEPSS rescore
  67. 2026-09-08 04:20 UTCEG score recompute
  68. 2026-09-08 04:20 UTCGHSA enrichment
  69. 2026-09-07 15:40 UTCGHSA enrichment
  70. 2026-09-07 02:59 UTCGHSA enrichment
  71. 2026-09-06 14:11 UTCEG score recompute
  72. 2026-09-06 14:11 UTCGHSA enrichment
  73. 2026-09-06 13:48 UTCEPSS rescore
  74. 2026-09-06 01:31 UTCEG score recompute
  75. 2026-09-06 01:31 UTCGHSA enrichment

Related CVEs(same CWE)

Frequently asked(5)

What is CVE-2026-72852?
CVE-2026-72852 is a high vulnerability published on August 20, 2026. hank-ai/darknet sizes a convolutional layer's weight and output heap buffers by multiplying configuration fields taken from a .cfg file in unchecked 32-bit int arithmetic. In src-lib/convolutionallayer.cpp, l.nweights is computed as (c / groups) * n * size * size and l.outputs as l.outh l.out_w…
When was CVE-2026-72852 disclosed?
CVE-2026-72852 was first published on August 20, 2026, with the most recent update on September 24, 2026. EchelonGraph re-ingests CVE updates from NVD on a 2-hour cycle, so this page reflects the latest published state.
Is CVE-2026-72852 actively exploited?
CVE-2026-72852 is not currently on CISA's Known Exploited Vulnerabilities catalog. FIRST EPSS estimates a 0.2% probability of exploitation in the next 30 days (10th percentile of EPSS-scored CVEs).
What is the CVSS score of CVE-2026-72852?
CVE-2026-72852 has a CVSS base score of 7.8 (a secondary CVSS source that NVD displays; NVD's own analysis pending).
How do I remediate CVE-2026-72852?
No fix for CVE-2026-72852 is confirmed yet. Until one is published, restrict network exposure of the affected system or apply the vendor's mitigation — for example, keep it off the internet or limit it to trusted networks — and watch the vendor's advisory for the fix. The vendor advisories EchelonGraph has for CVE-2026-72852 are linked in the Vendor Advisories panel on this page.

Dependency Blast Radius

Explore the affected products and dependency analysis for CVE-2026-72852

Explore →

Is Your Infrastructure Affected by CVE-2026-72852?

EchelonGraph automatically scans your cloud infrastructure and maps CVE exposure using blast radius analysis.

Book a Demo →Browse All CVEs