add-hook
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input (hook name, description, and interactive demo logic) to generate implementation files, test suites, and documentation across multiple files.
- Ingestion points: User-provided hook name and description in
SKILL.md(metadata) and file generation templates. - Boundary markers: None present to delimit user-provided content from agent instructions.
- Capability inventory: File-write operations in
src/registry/hooks/,skills/shadcn-hooks/references/, andskills/shadcn-hooks/SKILL.md; command execution viapnpm test. - Sanitization: The skill does not provide instructions for sanitizing or escaping user-provided text before interpolation into prompt-accessible files.
- [DYNAMIC_EXECUTION]: The skill generates source code and test files (
index.test.ts) based on user input and templates, then instructs the agent to execute them. - Evidence:
pnpm test run src/registry/hooks/use-<name>/index.test.ts - Context: While this is a standard verification workflow for a development-focused skill, it involves the execution of code generated dynamically during the session.
Audit Metadata