refactor
Pass
Audited by Gen Agent Trust Hub on Jun 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes various development-related shell commands such as "bun test", "npm test", "pytest", "tsc", and "eslint" to verify code changes during the refactoring process.
- [COMMAND_EXECUTION]: It utilizes the "sg" (ast-grep) CLI and a local Python script "scripts/ast_grep_helper.py" to automate code transformations.
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface. It ingests untrusted project data (source code) through commands like "cat", "grep", and LSP tools. This data is incorporated into the agent's context without sanitization or boundary markers. Since the skill has the ability to write to the filesystem and execute shell commands, malicious instructions embedded in the code being analyzed could theoretically influence the agent's output or actions.
Audit Metadata