omh-terminal

Pass

Audited by Gen Agent Trust Hub on Sep 25, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user requests to generate shell commands. It mitigates injection risks by establishing a clear 'prepared-vs-observed' evidence boundary and requiring the AI to treat prepared guidance as advisory rather than proof of execution.
  • Ingestion points: Processes user requests, target contexts, and delivery expectations.
  • Boundary markers: Utilizes command_result_manifest/v1 and explicit safety gates to delimit untrusted input from system state.
  • Capability inventory: Facilitates shell command preparation, environment configuration, and working directory management.
  • Sanitization: Enforces safety rules that require explicit human confirmation for destructive, network-facing, or credential-accessing commands.
  • [COMMAND_EXECUTION]: The skill is explicitly designed to supervise shell, CLI, and package-manager tasks. It includes a specific runtime recording command (omh runtime record) used for internal workflow tracking within the developer's framework.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 25, 2026, 08:34 PM
Security Audit — agent-trust-hub — omh-terminal