testing-chaos
Pass
Audited by Gen Agent Trust Hub on Feb 28, 2026
Risk Level: SAFECOMMAND_EXECUTIONNO_CODE
Full Analysis
- [SAFE]: The skill does not contain any malicious patterns such as prompt injection, hardcoded credentials, or unauthorized data exfiltration. All provided content is focused on the stated purpose of chaos engineering.
- [COMMAND_EXECUTION]: The SKILL.md file includes numerous examples of shell commands (
npm run test:chaos:*) that the agent is expected to execute. These commands are designed to perform disruptive operations like network failure injection and resource exhaustion; however, this behavior is consistent with the primary purpose of chaos testing and is documented for developers. - [NO_CODE]: Although the skill description claims to implement complex failure injection mechanisms (latency, service outages, resource pressure), the actual logic for these operations is absent from the provided repository. The only script included (
scripts/example.sh) is a generic boilerplate, and the skill relies on the presence of external npm scripts in the user's local environment.
Audit Metadata