Answer in brief
CVE-2025-6237 records a Critical severity (CVSS 9.8) vulnerability in InvokeAI has External Control of File Name or Path. The current sources do not mark it as known exploited. The current feed maps invokeai (pip), invokeai (pypi). Check affected ranges and fixed versions before updating.
Analysis pending evidence review
HOL Guard separates source facts from reviewed analysis. See the methodology.
CVSS is 9.8. The current sources do not mark it as known exploited. Treat this as a source-backed prioritization signal, not a statement about your environment.
Analysis status
Analysis pending evidence review
Factual feed record only; HOL analysis is not approved for indexing. Read the methodology.
The current feed maps invokeai (pip), invokeai (pypi). Check affected ranges and fixed versions before updating.
| Package | Affected range | Fixed version |
|---|---|---|
| invokeaipip | <6.7.0 | 6.7.0 |
| invokeaipypi | >=0 <6.7.0 | 6.7.0 |
Published upstream
Sep 18, 2025
Evidence: source:ghsa:source_dates:source-dates:recordSource modified
Oct 27, 2026
Evidence: source:ghsa:source_dates:source-dates:recordFirst seen by HOL
Oct 27, 2026
### Path Traversal Vulnerability in InvokeAI A path traversal vulnerability in **InvokeAI** (versions < 6.7.0) allows an unauthenticated remote attacker to read files outside the intended media directory via the **bulk downloads** API. The endpoint accepts a user-controlled file/item name and concatenates it into a filesystem path without proper canonicalization or allow-listing. By supplying sequences such as `../` (or absolute paths), an attacker can cause the server to traverse directories and return arbitrary files. In certain storage or back-end configurations, abusing attacker-controlled paths can also lead to unintended overwriting or deletion of files referenced by the crafted path. The issue is fixed in **6.7.0**, which normalizes and validates input paths and rejects traversal attempts. **Affected versions:** `< 6.7.0` **Patched version:** `6.7.0`
Quoted source text, attributed separately from HOL analysis.