Java CWE Security Skills Collection

Pass

Audited by Gen Agent Trust Hub on Mar 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill documentation provides various shell command patterns using grep (e.g., grep -rn "Runtime.getRuntime().exec") intended to help users and AI agents identify vulnerable code segments within a Java project. These are legitimate security auditing tools.
  • [CREDENTIALS_UNSAFE]: Instructional examples in several SKILL.md files (such as cwe-798-hardcoded-credentials and cwe-259-hardcoded-password) include hardcoded string literals for passwords and API keys (e.g., sk-1234567890abcdef). These are explicitly used as negative examples within 'Vulnerable Pattern' sections to teach AI agents how to recognize and remediate such insecure practices.
  • [EXTERNAL_DOWNLOADS]: The README.md and installation guides suggest the use of the npx skills command to download the skill collection from the author's GitHub repository, and provide links to the official MITRE CWE database for reference.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 6, 2026, 09:49 PM