agent-harness-fault-injection
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The file contains purely instructional Markdown content with no embedded executable scripts or commands.
- [SAFE]: The skill mandates strict safety boundaries, including the use of disposable sandboxes and the prohibition of testing against production data or live credentials.
- [SAFE]: Security best practices are explicitly integrated, such as requiring the redaction of secrets and personal data from test evidence and timelines.
Audit Metadata