The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes dynamic context injection (exclamation mark followed by backticks) to execute a shell command when the skill is loaded. This command extracts the catalog version and entry count from the CISA KEV JSON file to provide the agent with current context.
[EXTERNAL_DOWNLOADS]: The skill performs a network request to CISA's official website to fetch the Known Exploited Vulnerabilities catalog. This is a standard operation for security triage tools and targets an established authoritative source.
[PROMPT_INJECTION]: The skill includes a dedicated safety notice that instructs the AI agent to ignore any instructions embedded within vulnerability scan outputs, code comments, or external advisories. This provides a defensive layer against indirect prompt injection where an attacker might attempt to influence the triage result through malicious data fields.

cve-triage