Loading vulnerability…
CVE-2025-51867 is an insecure direct object reference (IDOR) vulnerability affecting Deepfiction AI (deepfiction.ai) through June 3, 2025. Based on the provided description, the issue allows an attacker to obtain sensitive information from the /browse/stories endpoint and use that information to interact with the platform's LLM using other users' credits. The vulnerable condition appears to stem from insufficient authorization checks on object references or identifiers exposed by the application, enabling an attacker to act on resources associated with other accounts rather than being restricted to their own.
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.
No valid public exploits. Mallory filtered out 1 candidate as fakes, detection scripts, or README-only repos.
All candidate exploits were filtered out by Mallory's validation.
2 sources tracked across advisories and community write-ups. News coverage will land here when it surfaces.
No news coverage yet. Advisories and community discussion only.
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.