GHSA-5XXX-QHH7-9287
2026-08-10
gitpython-developers/gitpython
Repo.blame contents reads arbitrary file
AI root cause · Injection & unsafe execution · Incomplete remediation
See root cause and fix →Loading…
Georgia Tech SSLab · security research
Vibe Security Radar
195 confirmed vulnerabilities where AI-written code introduced, enabled, or failed to close the flaw.
- Confirmed true positives
- 195 cases
- Covered advisories2025-05 – 2026-08
- 23,861 advisories6,431 reviewed · 17,430 under analysis
- Peak advisory month
- Jul 2026 · 43 cases
Most common AI tool- Claude · 150 of 195
Most represented project- openclaw/openclaw · 36 of 195
Disclosure trend
Disclosures over time
When these cases were publicly disclosed.
- Dated
- 195/195
- Range
- Sep 2025–Aug 2026
- Cadence
- Monthly
- Through
- 2026-08-16
0
0
2
1
0
18
30
30
31
28
43
9
Sep '25Nov '25Jan '26Mar '26May '26Jul '26
| Month | Findings |
|---|---|
| Sep 2025 | 0 |
| Oct 2025 | 0 |
| Nov 2025 | 2 |
| Dec 2025 | 1 |
| Jan 2026 | 0 |
| Feb 2026 | 18 |
| Mar 2026 | 30 |
| Apr 2026 | 30 |
| May 2026 | 31 |
| Jun 2026 | 28 |
| Jul 2026 | 43 |
| Aug 2026 | 9 |
* Aug 2026 is partial.
Patterns
Who, where, and why
AI tooling
Who appears on the candidate change
184 of 195 findings name a tool on every candidate commit.
- Claude150
- GitHub Copilot14
- ChatGPT/Codex11
- Cursor8
- claude-flow1
Base rate sibling
Share of AI-marked commits
23,977 AI-marked of 160,719 commits, same repos and window. Absolute case counts mean nothing without this denominator.
- Claude76.0% commits
cases 81.5% (150) · 1.07× cases vs commits
- GitHub Copilot18.9% commits
cases 7.6% (14) · 0.40× cases vs commits
- Cursor3.1% commits
cases 4.3% (8) · 1.40× cases vs commits
- claude-flow1.0% commits
cases 0.5% (1) · 0.53× cases vs commits
- ChatGPT/Codex1.0% commits
cases 6.0% (11) · 5.87× cases vs commits
Why some tools look over-represented
Claude self-marks nearly every commit; GPT/Codex and Cursor usually leave no marker, so their shares are floors and ratios above 1× are upper bounds, not proof a tool writes worse code. A ratio near 1× means cases track usage.
Root causes
What went wrong
One root-cause category per finding.
- Injection & unsafe execution51
- Authentication & access control39
- SSRF & network boundaries23
- Validation & fail-open logic22
- Path & link handling15
- Resource abuse & availability14
- Other / insufficient public mechanism detail31
Languages
Where the vulnerable code lives
Top languages; all remain available in Findings.
TypeScript29
JavaScript27
Python27
Go7
Rust6
C#5
Projects
Repositories with the most findings
Top repositories; filter the full index in Findings.
- openclaw/openclaw36
- mervinpraison/praisonai14
- czlonkowski/n8n-mcp8
- gitpython-developers/gitpython6
- jahlives/openssl_encrypt5
- coollabsio/coolify4
Featured findings
See the vulnerable code and fix
GHSA-WVPP-8HX9-P66J
2026-08-07
gitpython-developers/gitpython
The check_unsafe_options guard can be bypassed on every guarded method (clone/clone_from, fetch/pull/push, ls_remote, iter_commits, blame, archive) by combin...
AI root cause · Injection & unsafe execution · Direct introduction
See root cause and fix →CVE-2026-18980 (GHSA-CW23-QWR7-C655)
2026-08-06
nearai/ironclaw
A vulnerability was identified in nearai ironclaw up to 0.29.1. Affected is the function classify_command_risk of the file src/tools/builtin/shell.rs. Such m...
AI root cause · Injection & unsafe execution · Causal contribution
See root cause and fix →CVE-2026-50568 (GHSA-R5JH-Q2MW-GCX4)
2026-07-28
fission/fission
SanitizeFilePath in pkg/utils/utils.go validated that a path stayed under a safe directory by calling strings.HasPrefix(path, safedir). This is a lexical che...
AI-written code was flawed · Validation & fail-open logic · New attack surface
See root cause and fix →CVE-2026-59233 (GHSA-4FXP-2M36-QV64)
2026-08-10
roskus/prospero-flow-crm
Missing Authorization in the permission management component in Roskus Prospero Flow CRM before 5.2.1 allows any authenticated user to grant any role, includ...
AI root cause · Authentication & access control · Incomplete remediation
See root cause and fix →GHSA-3WXW-XV34-2FRG
2026-08-10
gitpython-developers/gitpython
TagReference positional file option
AI root cause · Path & link handling · Incomplete remediation
See root cause and fix →Media coverage
Q&A
- What counts as an AI-contributed vulnerability?
- The vulnerable behavior must depend on an AI-authored code change and be reversed by a real security fix.
- Is every AI-related commit shown?
- No. Each finding shows the contributing change and minimum fix, not unrelated search candidates.
- Is this the total number of AI security bugs?
- No. These 195 findings are confirmed true positives from public advisories whose Git history binds the flaw to an AI-authored change. They are a lower bound, not a census.
- Can a finding be corrected?
- Yes. Open a GitHub issue with the advisory, commit, or release evidence that changes the conclusion.