search-macos26-docs
Pass
Audited by Gen Agent Trust Hub on Jul 5, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill uses
subprocess.runandsubprocess.callto interact with the local macOS SDK toolchain, specificallyxcrunandswift-symbolgraph-extract. All calls are implemented using list-based argument passing, which is the standard secure practice to prevent shell injection vulnerabilities. - [DATA_EXFILTRATION]: Analysis of the Python scripts confirms that no network operations are performed. The skill strictly reads local SDK files and manages a local cache directory for symbol search results.
- [PROMPT_INJECTION]: The instructions provided in
SKILL.mdare purely technical and focus on the task of documentation search. There are no attempts to override agent behavior, bypass safety filters, or extract system prompts.
Audit Metadata