react-hook-form-zod
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONMETADATA_POISONING
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill templates define structures for ingesting external data from API responses and HTTP request bodies, which establishes a vulnerability surface for indirect prompt injection if data is processed by an agent. \n
- Ingestion points:
templates/async-validation.tsx(processes API fetch responses) andtemplates/server-validation.ts(processes incoming request JSON). \n - Boundary markers: The skill consistently promotes the use of Zod schemas to enforce strict data boundaries. \n
- Capability inventory: Includes shell script execution via
scripts/check-versions.shand network operations viafetch. \n - Sanitization: Demonstrates robust sanitization patterns using Zod's validation methods including regex, length constraints, and refinements. \n- [COMMAND_EXECUTION]: The skill includes a shell script designed for development environment auditing. \n
- Evidence:
scripts/check-versions.shexecutes thenpm viewcommand to retrieve package metadata from the public NPM registry. \n- [METADATA_POISONING]: A discrepancy exists between the authorship information provided in the plugin manifest and the skill's associated metadata. \n - Evidence: The
.claude-plugin/plugin.jsonfile identifies 'Jeremy Dawes' as the author, whereas platform-level context attributes the skill to 'fellipeutaka'.
Audit Metadata