cve-triage

Pass

Audited by Gen Agent Trust Hub on Jul 20, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill fetches the latest KEV catalog metadata from the official Cybersecurity and Infrastructure Security Agency (CISA) domain (cisa.gov). This is an informative step to provide the agent with the current version and release date of the vulnerability database.
  • [COMMAND_EXECUTION]: Uses a shell command at load time to parse the CISA KEV JSON file using Python. This command is restricted to extracting specific metadata fields (version, count, date) from the trusted source.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external data such as vulnerability scan results. It includes a 'Prompt Injection Safety Notice' that instructs the agent to ignore any embedded instructions within that data, such as attempts to artificially lower severity scores.
  • [REMOTE_CODE_EXECUTION]: A remote execution pattern is used to pipe a JSON feed from CISA to a Python script. This behavior is used solely for data parsing from a well-known service to populate version information.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 20, 2026, 02:11 PM
Security Audit — agent-trust-hub — cve-triage