Picklescan Bypass via Numpy Function Deserialization

picklescan before 0.0.33 fails to detect malicious pickle files that invoke numpy.f2py.crackfortran.myeval function through the reduce method. Attackers can craft malicious pickle files embedding arbitrary code that evades picklescan detection and executes remote code when loaded.

Casky was already ahead

This CVE exploits attack patterns that Casky's 446matched skills already investigate — long before this vulnerability was disclosed. Claude's reasoning model maps these techniques to MITRE ATT&CK, so practitioners who ran these skills have already seen the threat behaviour in their findings.

Analysis

CVE-2025-71365 exploits a detection gap in picklescan, a security tool designed to identify malicious serialized Python objects before they're deserialized. The vulnerability allows attackers to craft pickle files that invoke the numpy.f2py.crackfortran.myeval function through the reduce method—a technique that evades picklescan's inspection logic while maintaining the ability to execute arbitrary code when loaded. This is critical because pickle deserialization is a known attack vector in Python environments, particularly affecting data science and ML pipelines where pickle files are commonplace for model persistence. Organizations relying on picklescan as their primary defense against malicious serialized objects are directly impacted, as are any systems accepting untrusted pickle inputs from external sources.

Casky's 446 matching skills leverage Claude's extended reasoning to detect the attack patterns underlying this vulnerability across MITRE ATT&CK techniques TA0001 (Reconnaissance), TA0043 (Reconnaissance via OSINT), and TA0002 (Resource Development). Practitioners using Casky would identify detection gaps in deserialization validators (CWE-502 unsafe deserialization) by correlating behavioral indicators: unusual object reduction chains, unexpected numpy/scipy module imports, function invocation patterns that bypass allowlisting, and proof-of-concept pickle generation workflows. The platform's findings would highlight that attackers craft these bypasses during Resource Development phases, then exploit them in Initial Access scenarios where pickle files are loaded without secondary validation—enabling the reconnaissance and execution chains MITRE tracks. This vulnerability exemplifies why static scanning tools alone are insufficient; Casky's skill mapping helps teams understand the complete attack narrative and implement compensating controls like sandboxed deserialization and strict input validation beyond tool-based detection.

Live Threat IntelligenceComing soon

Casky Risk Score

—

EPSS Probability

—

Shodan Exposure

—

GreyNoise Signal

—

Composite risk scoring from EPSS, CISA KEV, Shodan, and GreyNoise — 21 security APIs correlated into a single Casky Risk Score. Coming in Casky Pro. Join early access →

446 Casky skills cover this attack pattern

These skills use Claude AI's reasoning model to surface findings in the same attack categories as CVE-2025-71365.

Abuse of Public-Facing API: Mobile API Abuse

mobile security · low

Run

abusing-dpapi-for-credential-access

red teaming · high

Run

abusing-shadow-credentials-for-privesc

red teaming · high

Run

MITRE ATT&CK Techniques

TA0001 TA0043 TA0002

Investigate the attack patterns behind CVE-2025-71365

Casky has 446 skills that investigate the attack patterns behind CVE-2025-71365. Run one and get CVSS-scored findings in 3 minutes.

Run the skill that detects this →

X Instagram LinkedIn