sanity-best-practices
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references numerous official and community-supported packages from the npm registry (e.g.,
@sanity/client,next-sanity,hydrogen-sanity,@portabletext/react) used for integrating Sanity with various frontend frameworks. - [COMMAND_EXECUTION]: The instructions guide the agent to use the official Sanity CLI (
npx sanity) for routine development tasks such as project initialization, schema deployment, and infrastructure management. These commands are standard for the Sanity ecosystem. - [INDIRECT_PROMPT_INJECTION]: The skill documents methods for fetching content from the Sanity Content Lake. This creates an attack surface where untrusted data from the CMS could be ingested into the agent's context.
- Ingestion points:
client.fetch()calls are documented throughout the reference files (e.g.,nextjs.md,groq.md) to pull data into application components. - Boundary markers: The skill recommends using
defineQueryfor type safety, though it does not explicitly enforce security boundaries for the data content itself. - Capability inventory: The agent is granted the ability to execute CLI commands and write code, which could be exploited if malicious content is processed without care.
- Sanitization: The skill emphasizes the use of Sanity's validation rules in schema definitions to ensure data integrity and provides patterns for cleaning stega markers from logic-sensitive strings.
Audit Metadata