test-attack-technique

Pass

Audited by Gen Agent Trust Hub on Feb 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local Go code using go run and cloud provider CLI tools (AWS, Azure, GCP, Kubernetes) to manage infrastructure and simulate attacks.
  • [DATA_EXFILTRATION]: The skill instructions suggest capturing command output which may contain sensitive information, such as secret values from AWS Secrets Manager, and including them in unencrypted local HTML reports. This creates a data exposure risk on the local file system.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection by parsing untrusted technique source code and command outputs to generate its execution plan and reports.
  • Ingestion points: Technique source code in the v2/ directory and unstructured command output from the stratus tool.
  • Boundary markers: None implemented in the instructions for parsing external data.
  • Capability inventory: Execution of arbitrary Go code, execution of cloud provider CLI tools, and file system write operations.
  • Sanitization: No sanitization or validation of technique-specific content is performed before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 25, 2026, 10:09 AM