rebase
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill facilitates the execution of shell commands using standard version control binaries, including
gitandsl(Sapling), to perform history-rewriting operations such as rebasing, fetching, and stashing. - [INDIRECT_PROMPT_INJECTION]: The skill accepts an optional branch or bookmark name from the user, which is subsequently interpolated into shell commands (e.g.,
git rebase <target>). This creates a surface for potential command injection if the input is malicious. - Ingestion points: User-provided argument representing a branch or bookmark name.
- Boundary markers: The skill does not define explicit boundary markers or instructions for the agent to validate the input string.
- Capability inventory: The skill utilizes shell execution capabilities to run
gitandslcommands. - Sanitization: No explicit sanitization logic or regex validation is provided for the interpolated variable.
Audit Metadata