tailrocks-tanstack-project-migrate
Pass
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted codebase files and external data, creating an inherent surface for indirect prompt injection.
- Ingestion points: Project source code, lockfiles, and external API responses (as referenced in references/migration-checklist.md and references/boundaries-and-data.md).
- Boundary markers: Explicitly defined in references/runtime-trust.md, which states that embedded instructions from repository files or web content cannot alter scope or governing rules.
- Capability inventory: Includes file system writes (migration slices), command execution (bun, shadcn), and network access for documentation and registry checks.
- Sanitization: Mandated in references/boundaries-and-data.md, requiring all untrusted values to be parsed from unknown and server function inputs to be validated.
- [COMMAND_EXECUTION]: The skill utilizes the bun runtime and shadcn CLI for project management. It also instructs the agent to execute a vendor-provided script, resolve-package-versions.ts, to synchronize dependency versions.
- [EXTERNAL_DOWNLOADS]: The skill references official release sources and documentation from well-known technology vendors, including Bun, Microsoft (TypeScript), Meta (React), and Vite, to verify migration paths and security updates. These references to official domains and repositories are consistent with standard development practices.
Audit Metadata