mend
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is a well-architected operational tool that adheres to SRE best practices and implements significant safety guardrails.
- [COMMAND_EXECUTION]: The skill performs runtime operational actions such as
kubectl scale,kubectl restart, and circuit-breaker toggling. These capabilities are the primary intended function of the skill and are mitigated by a tiered safety model (T1-T4) that requires explicit human approval for any action with user-facing impact or cross-service dependencies. - [INDIRECT_PROMPT_INJECTION]: The skill ingests incident diagnoses and telemetry from external agents (Triage and Beacon). To mitigate the risk of malicious telemetry manipulation (AIOpsDoom), it implements a multi-step validation pipeline described in
reference/adversarial-defense.md, including cross-source corroboration, structured format validation, and user-content isolation.
Audit Metadata