spree-api-v3

Pass

Audited by Gen Agent Trust Hub on Jul 7, 2026

Risk Level: SAFECREDENTIALS_UNSAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The file SKILL.md contains a hardcoded publishable API key (pk_CzEKBTWFiuNLgz4wciLsS59n) within a curl example for the Store API. While publishable keys are often intended for public use, hardcoding specific credential values in documentation is a security concern.
  • [COMMAND_EXECUTION]: The skill provides numerous curl command examples for interacting with the Spree API, which involve network operations targeting external domains like example.com.
  • [EXTERNAL_DOWNLOADS]: The skill references several official vendor-provided software development kits and documentation packages, specifically @spree/sdk, @spree/admin-sdk, and @spree/docs.
  • [PROMPT_INJECTION]: The skill documents an indirect prompt injection surface by instructing the agent on how to ingest and parse data from external API responses.
  • Ingestion points: API response data from /api/v3/store/* and /api/v3/admin/* endpoints as described in the documentation.
  • Boundary markers: No specific delimiters or instructions to ignore potential instructions embedded within the API response data are present.
  • Capability inventory: The skill leverages curl for network requests and automated processing of the resulting JSON data.
  • Sanitization: The documentation highlights server-side sanitization through Ransack whitelisting of attributes and associations.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 7, 2026, 05:16 PM
Security Audit — agent-trust-hub — spree-api-v3