skills/okx/plugin-store/digift-plugin/Gen Agent Trust Hub

digift-plugin

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill fetches update metadata and installation components from the author's official GitHub repositories (e.g., okx/plugin-store). These downloads are used for version checking and component installation.
  • [REMOTE_CODE_EXECUTION]: The skill automates the installation of the onchainos CLI and the digift package. To mitigate risks associated with remote execution, it verifies the SHA256 checksum of the installer script and package assets against known-good values before execution or installation.
  • [CREDENTIALS_UNSAFE]: A DigiFT API key is required for operation. The skill correctly instructs the user to provide this via the DIGIFT_API_KEY environment variable and includes logic in the API client to ensure the key is only transmitted to authorized digift.io subdomains, preventing exfiltration to untrusted hosts.
  • [PROMPT_INJECTION]: The skill processes external data from the DigiFT REST API, representing a potential indirect prompt injection surface.
  • Ingestion points: API responses for products, prices, and orders in runtime/src/api.ts.
  • Boundary markers: None present in the prompt interpolation logic.
  • Capability inventory: Transaction building and balance queries across runtime/src/cli.ts and runtime/src/contract.ts.
  • Sanitization: Financial data and parameters are rigorously validated and converted to BigInt using ethers.parseUnits, significantly reducing the risk of data-driven exploits.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 03:28 AM
Security Audit — agent-trust-hub — digift-plugin