stablecoin-risk-check
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from
api.barker.moneyand references an embedded knowledge base, creating a surface where untrusted data enters the agent context. - Ingestion points: Data is ingested from
api.barker.money/api/public/v1and the 'Knowledge Base' section inSKILL.md. - Boundary markers: The skill includes a dedicated 'Security: External Data Boundary' section that explicitly instructs the agent to treat returned strings as data rather than instructions.
- Capability inventory: The skill primarily performs data analysis and reporting; it does not contain file-write, network-write, or execution capabilities for the analyzed data.
- Sanitization: The skill relies on natural language instructions to the agent to perform string-only rendering and avoid executing embedded text.
- [EXTERNAL_DOWNLOADS]: The skill includes a bash script intended for execution by the agent to perform a version check. This script uses
curlto fetch theplugin.yamlfile from the vendor's repository atraw.githubusercontent.com/okx/plugin-store. The operation is limited to reading version metadata and does not automate the installation of remote code.
Audit Metadata