omh-deploy-and-monitor
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill documentation includes examples for using the
omhcommand-line utility, a vendor-specific tool for the 'Oh My Hermes' platform, to record workflow transitions and deployment status. - Evidence: The command
omh runtime record --skill deploy-and-monitor --harness app-delivery-loop --status startedis documented in the runtime evidence section ofSKILL.md. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-provided release metadata, which represents a potential surface for indirect prompt injection if the inputs contain malicious instructions.
- Ingestion points: The skill collects 'release scope', 'environment', 'health signals', and 'rollback owner' as required inputs.
- Boundary markers: There are no explicit delimiters or boundary instructions defined for the interpolation of these user-provided fields into status records or checklist outputs.
- Capability inventory: The skill has the capability to execute
omhtool commands and delegate tasks to platform-native subagents. - Sanitization: No explicit validation, escaping, or sanitization requirements are mentioned for the ingested metadata before it is processed by the agent.
- [SAFE]: The skill incorporates strong safety constraints that mandate observed evidence before reporting deployment success and strictly separates standard release monitoring from active incident response protocols.
Audit Metadata