cspm-casb-attack

Fail

Audited by Socket on Sep 19, 2026

3 alerts found:

MalwareSecurityAnomaly
MalwareHIGH
SKILL.md

MALICIOUS. This skill’s actual footprint is fundamentally offensive: it teaches an AI agent to steal tokens, read local credential stores, suppress security findings, evade CASB controls, tamper with IaC state, and maintain stealth in real environments. The data flows and capabilities are incompatible with a benign developer-assistance skill and create high risk of unauthorized access, credential abuse, and exfiltration.

Confidence: 97%Severity: 96%
SecurityMEDIUM
guides/quick-reference-card.md

The artifact is a plain-text offensive cloud-security playbook. It does not itself execute or install malware, but it contains highly actionable instructions for CASB evasion, OAuth token replay, direct-to-origin access, policy bypass, Terraform state poisoning, credential extraction, and CSPM alert suppression. It should be treated as dangerous offensive content and restricted to explicitly authorized security testing with strong credential, state-file, and change-control safeguards.

Confidence: 99%Severity: 96%
AnomalyLOW
test-cases.md

The fragment documents controlled cloud-security attack simulations and lab setup rather than containing apparent malware. It includes operationally risky commands that can retrieve instance-role credentials, enumerate S3, alter EC2 metadata settings, and intercept TLS traffic if run against real or unintended systems. Use only isolated, authorized AWS resources and local endpoints. The visible content shows no clear malicious package behavior, but it has meaningful dual-use security risk.

Confidence: 96%Severity: 55%
Audit Metadata
Analyzed At
Sep 19, 2026, 04:41 PM
Package URL
pkg:socket/skills-sh/brucesongs%2Fkali-claw%2Fcspm-casb-attack%2F@a516313eb75b5683f3ba46f1d616543f9316413b74a4cdd77d28220b21f842b5