iterate-structure-outline
Pass
Audited by Gen Agent Trust Hub on Oct 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external artifacts such as ticket files and user feedback, which may contain indirect instructions. Ingestion points occur in Steps 1, 2, and 3 where artifacts and feedback are read into context. Capabilities include file writing and bash commands. As a mitigation, Step 3 explicitly instructs the agent to verify information via research sub-agents rather than accepting input blindly.
- [COMMAND_EXECUTION]: Shell commands are used for filesystem and repository state inspection. Evidence includes the use of 'ls -La' to list task artifacts and 'git rev-parse' to identify worktrees. These are restricted, benign operations within the developer environment.
Audit Metadata