oa-dispatch
Pass
Audited by Gen Agent Trust Hub on Mar 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill centers on the use of the oa CLI to dispatch tasks to external backends.\n- [EXTERNAL_DOWNLOADS]: Documentation instructs users to install the oa utility via 'go install github.com/1broseidon/oneagent/cmd/oa@latest'. As this resource belongs to the skill author, it is considered a vendor-provided tool.\n- [REMOTE_CODE_EXECUTION]: Dispatched agents are tasked with modifying source code and executing verification commands, allowing for the execution of code generated by external models on the local environment.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its core functionality of processing local data.\n
- Ingestion points: External data enters the context through files specified in dispatch prompts (e.g., internal/auth/handler.go).\n
- Boundary markers: Absent; there are no instructions for using delimiters to separate instructions from target file content.\n
- Capability inventory: The oa tool enables filesystem writes and shell command execution.\n
- Sanitization: Absent; file contents and prompts are passed to model backends without validation or escaping.
Audit Metadata