web-validation-zod

Pass

Audited by Gen Agent Trust Hub on Aug 14, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides standard, non-malicious coding patterns for data validation using the Zod library.
  • [INDIRECT_PROMPT_INJECTION]: The skill involves reading project files to maintain consistency, creating a surface for potential indirect prompt injection.
  • Ingestion points: The skill instructs the agent to read existing files within the **/schemas/** directory.
  • Boundary markers: No explicit delimiters or instructions to ignore embedded prompts are provided for the ingestion of schema files.
  • Capability inventory: The skill is limited to code generation; it does not request network access, sensitive file writes, or shell command execution.
  • Sanitization: No sanitization is performed on the ingested schema data, but the final output is code intended for developer review.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 14, 2026, 03:50 PM
Security Audit — agent-trust-hub — web-validation-zod