qa-resilience-fault-injection

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill is designed to orchestrate the execution of fault injection commands via a runner agent. It specifically utilizes tools such as Toxiproxy for network manipulation, Pumba for container management (kill/pause), and stress-ng for resource exhaustion. While these are powerful commands, they are scoped to non-production environments.
  • [EXTERNAL_DOWNLOADS]: The skill references several reputable open-source projects for chaos engineering, including Shopify's Toxiproxy and the Pumba container testing tool. These references point to official GitHub repositories and are used to provide standard tooling for resilience testing.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing user-provided target environments and experiment parameters. It implements mitigation strategies through a 'Double-Gate' workflow, requiring explicit human approval before any injection occurs, and strict environment isolation checks to identify and block production-level targets.
  • [SAFE]: The skill demonstrates high security maturity for its use case by implementing 'Red-Line' protections. These include mandatory steady-state baseline measurements, bounded blast-radius constraints for all faults, and an 'always rollback' policy to ensure the testing environment is returned to its original state regardless of the experiment's outcome.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 09:28 AM