auditing-appstore-readiness
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill reads and summarizes content from untrusted repositories, creating a surface for embedded instructions to influence the agent's audit report.
- Ingestion points: The
audit.mjsscript ingests data frompackage.json,app.json,Info.plist, and other project metadata files. - Boundary markers: The audit script encapsulates repo evidence within Markdown code blocks in the final report.
- Capability inventory: The agent has access to the
exectool, and theaudit.mjsscript executes subprocesses likegitandxcodebuild. - Sanitization: The script uses JSON parsing and regular expressions for data extraction, but the AI interprets the final text output.
- [COMMAND_EXECUTION]: The skill facilitates the execution of local development tools (
git,node,xcodebuild). The includedaudit.mjsscript usesspawnSyncwithshell: false, which provides a secure way to handle command arguments and reduces the risk of shell injection.
Audit Metadata