Loading vulnerability…
CVE-2026-53629 is a blind SQL injection vulnerability in GLPI affecting the history log filter functionality, referred to as LogBleed. The vulnerable code path is associated with Log::convertFiltersValuesToSqlCriteria() and SQL criteria handling in DBmysqlIterator::analyseCrit(). Attacker-controlled input supplied through the affected_fields filter can be incorporated into SQL logic without sufficient neutralization, enabling time-based blind SQL injection. Available information indicates the issue was reproducible on GLPI 10.0.25 and rejected in GLPI 10.0.26, indicating the flaw was fixed in that release.
Mallory correlates every CVE against your assets, your vendors, and active adversary campaigns. Know which vulnerabilities matter for you, not just which ones are loud.
What it means. What to do now. Patch path, mitigations, and the assume-compromise checklist.
What an attacker gets, and what they’ve been doing with it.
If you can’t patch tonight, do this now.
Patch, then assume compromise.
1 valid exploit after Mallory filtered fakes, detection scripts, and README-only repos.
Small standalone Python exploit repository for CVE-2026-53629 affecting GLPI's log/history filter. The repo contains one executable script (glpi_sqli_extract.py), a README explaining the vulnerability and usage, and a minimal .gitignore. The exploit is not framework-based. The script performs an authenticated web attack against GLPI by first requesting /front/login.php, extracting the _glpi_csrf_token from the HTML, and submitting credentials to the same endpoint. It then sends crafted GET requests to /front/log/export.php with itemtype=Entity, id=0, and filter[affected_fields][0]=OR::<payload>. This abuses the affected_fields parsing flaw described in the README to inject raw SQL into the WHERE clause. Capability-wise, the exploit is a time-based blind SQLi extractor. It verifies the bug with a simple sleep payload, then uses CASE WHEN ... THEN sleep() ELSE 0 END conditions and response timing to evaluate booleans. It determines the target string length and extracts characters one by one using ORD(MID(...)) with a binary search over printable ASCII. By default it reads SELECT password FROM glpi_users WHERE id=<user_id>, yielding the selected user's password hash. The README further states the same primitive can be adapted to extract API tokens, personal tokens, and SMTP/LDAP credentials from configuration tables. Operationally, the exploit requires valid credentials for an account with the GLPI 'logs' READ right and a vulnerable GLPI version. It is more than a detector because it automates data extraction, but it is still a focused PoC/operational script rather than a generalized framework module.
Products and vendors Mallory has correlated with this vulnerability. Open in Mallory to drill down to specific CPE configurations and version ranges.
Vendor-confirmed product mapping. Mallory continuously reconciles this list against your asset inventory.
4 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.
Query your assets running an affected version, and investigate the blast radius.
Every observed campaign linking this CVE to a named adversary.
Malware families riding this exploit, with evidence and IOCs.
YARA, Sigma, Snort, and vendor rules, auto-deployed to your SIEM.
Cross-references every affected SKU, including bundled OEM variants.
Community discussion across Reddit, Mastodon, and other social sources.