gsd-quick-batch
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes the
gsd_runCLI tool via a shell environment to perform argument parsing and validation. The instructions explicitly mandate quoting the input to prevent shell metacharacter expansion. - [INDIRECT_PROMPT_INJECTION]: The skill processes arbitrary task descriptions supplied by the user which are later dispatched to other agent instances.
- Ingestion points: User-supplied task lists in the
$ARGUMENTSvariable or within files referenced by the--fileargument inSKILL.md. - Boundary markers: The documentation requires wrapping untrusted task data in
DATA_START/DATA_ENDmarkers and using<security_context>blocks to signal to sub-agents that the content is data rather than instructions. - Capability inventory: Access to
Bashfor workflow coordination and theAgenttool for task dispatching. - Sanitization: The skill relies on a pre-defined CLI grammar parser (
quick-batch parse-args) to ensure arguments conform to expected formats before any tasks are created or dispatched.
Audit Metadata