crowdstrike-security
Fail
Audited by Gen Agent Trust Hub on Mar 31, 2026
Risk Level: CRITICAL
Full Analysis
- [SAFE]: The skill functions as a professional and legitimate resource for CrowdStrike Falcon platform experts, containing dense technical guidance on platform architecture and event search syntax.
- [COMMAND_EXECUTION]: The skill documents administrative commands such as kill, rm, and reg delete within the context of the Falcon Real Time Response (RTR) tool. These are standard operations for the defined expert persona and are governed by explicit boundary instructions requiring human confirmation before execution.
- [EXTERNAL_DOWNLOADS]: The documentation includes references to malicious URL placeholders, specifically identified as examples for sandbox submission and threat indicator extraction. These are instructional in nature and do not involve unauthorized network operations by the agent.
- [PROMPT_INJECTION]: System prompts and persona definitions are standard for a professional technical expert and do not attempt to bypass or override base AI safety guidelines.
Recommendations
- Contains 1 malicious URL(s) - DO NOT USE
Audit Metadata