omniroute-cli-cloud
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of instructions from local files via the
--prompt-fileflag, which are then transmitted to cloud agents like OpenAI Codex, Devin, or Jules. This represents an attack surface where a malicious local file could inject instructions to manipulate the behavior of the cloud agents. \n - Ingestion points: Usage of
--prompt-file ./task-description.mdin task creation commands. \n - Boundary markers: No explicit delimiters or warnings to ignore embedded instructions are provided in the documented usage. \n
- Capability inventory: The skill uses network operations to send data to external agent APIs. \n
- Sanitization: There is no evidence of filtering or validation of the input file content before it is processed by the agents. \n- [EXTERNAL_DOWNLOADS]: The documentation references an external project hosted on a third-party GitHub repository (
diegosouzapw/OmniRoute) as a requirement for the skill to function. \n - Evidence: Link to
https://raw.githubusercontent.com/diegosouzapw/OmniRoute/main/skills/omniroute-cli/SKILL.md. \n- [COMMAND_EXECUTION]: The skill's primary functionality is delivered through the execution of a third-party CLI tool (omniroute) for authentication and remote task management.
Audit Metadata