shadcn-component-review
Pass
Audited by Gen Agent Trust Hub on Jul 11, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection because its primary function is to ingest and process untrusted code provided by users for review.
- Ingestion points: Component code and layouts shared by users for auditing (found in SKILL.md).
- Boundary markers: The instructions lack specific delimiters or directives to the agent to treat comments or strings within the user code as non-executable data.
- Capability inventory: The skill allows the agent to execute shell commands (npx, grep) and interact with the project filesystem (found in SKILL.md and references/review-checklist.md).
- Sanitization: There is no process for filtering or escaping instructions that might be embedded in the components being analyzed.
- [COMMAND_EXECUTION]: The skill utilizes local shell commands to retrieve project configuration and perform text-based searches for audit violations.
- Evidence: Executes npx shadcn@latest info --json to gather framework details and uses grep to identify hardcoded color tokens in the codebase.
- [EXTERNAL_DOWNLOADS]: The skill references the use of npx to run the shadcn CLI, which facilitates fetching documentation and managing component configurations from external registries.
- Evidence: SKILL.md and references/theme-styles.md contain instructions to run npx shadcn@latest for project info and documentation retrieval.
Audit Metadata