The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process potentially malicious data such as system logs and process artifacts during incident triage. \n- Ingestion points: System logs, network connections, and process trees analyzed in Steps 3 and 4 of SKILL.md. \n- Boundary markers: The skill includes a 'Boundaries' section that explicitly prohibits evidence tampering and instructs the agent to focus on defense. \n- Capability inventory: The skill has access to Bash, Read, Write, and WebSearch tools for operational tasks. \n- Sanitization: Mitigation is provided through the use of a strict 'Output Format' markdown template to structure the agent's findings and recommendations. \n- [SAFE]: The skill implements security best practices for incident response, focusing on containment and evidence preservation without any detected malicious intent or exfiltration patterns.

incident-triage