OSV 1.4.0 · github-reviewed · 修改于 2026-08-26 01:28
发布时间
2026-08-26 01:28
GitHub 审查时间
2026-08-26 01:28
NVD 发布时间
—
源文件
advisories/github-reviewed/2026/08/GHSA-88g4-74f3-63x9/GHSA-88g4-74f3-63x9.json
There is an authenticated path traversal condition in the PDF export functionality, specifically within the PDF image resolution logic.
The issue may allow a privileged FAQ editor to cause the application to attempt reading files outside the intended content directory during PDF generation.
File:
src/phpMyFAQ/Export/Pdf/Wrapper.php
Function:
concatenatePaths()
The path resolution logic relies on locating the substring "content" within a user-controlled path:
$pos = strpos($trimmedFile, 'content');
$relativePath = substr($trimmedFile, (int) $pos);
If "content" is not present, strpos() returns false, which becomes 0 when cast to an integer.
As a result, the entire attacker-controlled path is preserved.
Example:
$trimmedFile = "../../../etc/passwd";
$pos = false;
$relativePath = "../../../etc/passwd";
The resulting path is later processed by:
file_get_contents($resolvedPath);
without canonicalization or a root-directory containment check.
FAQ Content
->
PDF Export
->
WriteHTML()
->
Wrapper::Image()
->
concatenatePaths()
->
file_get_contents()
Based on code review, a user with FAQ editing privileges may be able to store HTML containing crafted image paths that are processed during PDF generation.
Potential consequences may include:
This issue was initially detected by an internally developed SAST tool during analysis of the phpMyFAQ source code.
The finding was then manually investigated and validated through code review.
While the original scanner output classified the issue as a generic path traversal/local file inclusion pattern, manual analysis identified the specific root cause in the path resolution logic of concatenatePaths().
The following scenario is based on code review and intended to illustrate the potential impact:
A user with FAQ editing privileges creates or modifies a FAQ entry.
The FAQ content contains an HTML image tag with a crafted relative path that does not include the expected content directory reference.
The HTML content is stored and later processed by the PDF export functionality.
When a user requests the PDF version of the FAQ, the application invokes WriteHTML(), which eventually reaches Wrapper::Image().
concatenatePaths() constructs a filesystem path without canonicalization or directory containment validation.
The resulting path reaches file_get_contents(), causing the application to attempt reading a file outside the intended content directory.
Depending on sanitization behavior, file permissions, image validation, and PDF rendering behavior, the contents of the targeted file may potentially be exposed to the PDF consumer.
Based on my current analysis, exploitation would require a user capable of editing FAQ content and is therefore not considered an anonymous or unauthenticated attack vector.
Consider replacing substring-based path anchoring with:
realpath() canonicalization