omh-deliverable-package

Pass

Audited by Gen Agent Trust Hub on Sep 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill invokes the omh command-line utility to record the runtime status of the workflow (e.g., omh runtime record --skill deliverable-package --harness deliverable-package --status started). This is a standard operation within the intended framework environment.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data, specifically user requests and target contexts, which introduces a potential surface for indirect prompt injection.
  • Ingestion points: The skill ingests data from fields such as user request and target context as specified in the Required Inputs section of SKILL.md.
  • Boundary markers: The instructions explicitly require the agent to maintain a "prepared-vs-observed boundary" and to use "wrapper actions" to delimit workflow orchestration from actual evidence.
  • Capability inventory: The skill utilizes a specific subprocess call to the omh CLI tool for status recording within SKILL.md.
  • Sanitization: The skill instructs the agent to "expose missing tools, credentials, targets, or observations as user-visible gaps" rather than blindly obeying instructions within the processed data.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 25, 2026, 08:33 PM
Security Audit — agent-trust-hub — omh-deliverable-package