init-app-stack
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill performs legitimate scaffolding operations as described in its metadata. No malicious code or hidden instructions were found.- [COMMAND_EXECUTION]: The scaffolding script
scripts/create.pyruns several setup commands (e.g.,bun install,uv add). These are invoked using Python'ssubprocess.runwith a list of arguments, which is a secure method that avoids shell injection risks.- [EXTERNAL_DOWNLOADS]: The skill downloads packages and tools from standard registries and official vendor repositories (e.g.,github.com/bmsuisse/*). These downloads are necessary for the project setup and target trusted sources.- [INDIRECT_PROMPT_INJECTION]: The skill creates an injection surface by placing user-supplied project names into generated files likeREADME.md. \n - Ingestion points: Project name from
sys.argv[1]inscripts/create.py.\n - Boundary markers: Absent in the generated
README.mdheader.\n - Capability inventory: The resulting project enables file system operations and execution of generated scripts.\n
- Sanitization: The name is slugified for the project directory but used directly as a markdown header in the README.
Audit Metadata