sync-tree
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses Git commands such as
git push,git merge, andgit update-refto perform state changes in local repositories. While core to its function, these are high-capability operations. - [INDIRECT_PROMPT_INJECTION]: The skill extracts information from external sources to guide rebase conflict resolution.
- Ingestion points: Git commit messages and metadata from pull requests, issues, or tickets referenced in
references/rebase.md. - Boundary markers: No specific delimiters or 'ignore' warnings are used when processing this external data; the agent is instead told to establish 'intent' from them.
- Capability inventory: Write access to local Git branches and the execution of project-defined validation scripts.
- Sanitization: No sanitization or escaping of external content is specified.
- [DYNAMIC_EXECUTION]: During validation, the skill is instructed to discover and run project-specific checks from instructions and configuration files within the repository, which may lead to the execution of arbitrary code found in the project's local environment.
Audit Metadata