vitest-migration
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides the agent with access to a broad set of shell commands through the
Bashtool, includingnx,pnpm,git,node,sed,grep, andgh. These are used to survey the codebase, run tests, and perform codemod operations. - [DYNAMIC_EXECUTION]: The skill instructions involve advanced Node.js environment manipulation, specifically patching
Module._resolveFilenameandModule._loadwithintools/vitest/setup.mtsto handle custom module resolution and mocking. It also uses@swc-node/registerfor runtime TypeScript compilation and incorporates a write-guard throughexecArgvto monitor filesystem operations during test execution. - [INDIRECT_PROMPT_INJECTION]: The skill represents an attack surface for indirect injection as it ingests and processes untrusted data from repository files (source code, Jest/Vitest configurations, TSConfigs) and external sources such as
nx-cloudsandbox reports and GitHub PR data. The agent possesses high-privilege capabilities (file writing and shell execution) to act upon this data without explicit instruction boundary markers or sanitization steps mentioned in the workflow.
Audit Metadata