create-agent-with-sanity-context
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill provides documentation, architectural patterns, and reference code for building AI-powered chatbots using Sanity.io as a CMS backend.
- [EXTERNAL_DOWNLOADS]: The skill references official Sanity and AI SDK packages (e.g.,
@sanity/client,@ai-sdk/anthropic,ai) which are well-known and trusted vendor libraries. All external URLs point to official Sanity domains (sanity.io,api.sanity.io) or established documentation/provider sites. - [DATA_EXPOSURE_AND_EXFILTRATION]: While the skill involves handling API tokens (Sanity tokens, LLM keys), it correctly follows security best practices by recommending server-side storage in environment variables (
.env) and explicitly warns against exposing tokens to the browser. - [COMMAND_EXECUTION]: Shell commands provided in the documentation (e.g.,
npx sanity tokens add,curl) are standard administrative and testing commands intended for manual execution by the developer during setup. - [INDIRECT_PROMPT_INJECTION]: The skill implements an architecture that processes external data (Sanity content). However, it uses structured data fetching and recommends specific boundary markers in system prompts to mitigate injection risks. Findings in this category are assessed as safe due to the implementation of best practices for schema-aware querying.
Audit Metadata