omh-lifecycle-growth
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or unauthorized external communications were detected. The skill focuses on strategic planning and metadata-only record generation within the Hermes ecosystem.
- [COMMAND_EXECUTION]: The skill utilizes a local runtime tool
omhto process inputs and generate metadata. This is part of the expected functional behavior for the oh-my-hermes workflow and occurs in a controlled environment. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user inputs such as lifecycle objectives and audience segments. It provides an extensive 'fail-closed' policy that requires the agent to halt (HOLD) if critical inputs like consent or suppression are unknown, mitigating the risk of instructions being injected via external data.
- [DATA_EXFILTRATION]: Explicit safety rules prohibit the storage of user identity, event payloads, message bodies, or secrets in durable artifacts. This significantly reduces the attack surface for data exposure or exfiltration.
Audit Metadata