speckit-specify-zh
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes Git commands (git fetch, git branch, git checkout) and local scripts to manage the development workflow. These operations are standard for repository management tools and include sanitization for branch names derived from user input.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-supplied functional descriptions to generate requirements and user stories in spec.md. This is a core functionality of the skill; while it represents an attack surface for processing untrusted data, it is managed through a structured template system.
- [DYNAMIC_EXECUTION]: Provided Bash and PowerShell scripts use variable interpolation for file path and environment management. The scripts include logic to filter and sanitize these inputs (e.g., using sed to allow only alphanumeric characters in branch names) to mitigate injection risks.
- [SAFE]: The skill operates entirely within the local repository environment, populating a .specify directory with templates and scripts. No external downloads, data exfiltration, or persistence mechanisms were found.
Audit Metadata