Loading vulnerability…
PraisonAI versions through 4.6.33 expose MCP file-handling tools that accept attacker-controlled path values without enforcing containment within the intended rules directory. The MCP JSON-RPC dispatcher forwards caller-supplied tool arguments directly to handlers without enforcing their declared input schema. An attacker can use directory-traversal sequences in a rule-name argument to escape the designated directory and create or modify files writable by the account running the PraisonAI process. Writing a Python path-configuration file into a user-controlled Python package location can cause attacker-controlled code to execute when the user subsequently starts a Python process. The initial remediation was released in 4.6.34; later versions prior to 4.6.40 retained related incomplete path-handling and dispatcher-validation weaknesses.
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 public exploits tracked yet. Mallory keeps watching.
No public exploit code observed for this vulnerability.
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.
11 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.
An incomplete fix for a PraisonAI file-read/path handling vulnerability that allowed unauthenticated MCP tool calls to read arbitrary files accessible to the host user, due to unchanged vulnerable handlers and unvalidated dispatcher arguments prior to version 4.6.40.
A path traversal vulnerability in PraisonAI's MCP functionality that was actively weaponized in observed attacks.
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.