xl1-testing

Pass

Audited by Gen Agent Trust Hub on Nov 5, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources such as blockchain state and datalake payloads, which could potentially contain malicious instructions targeting the agent.
  • Ingestion points: Data is ingested through connection.viewer for chain state and RestDataLakeViewer.get for off-chain payload bodies.
  • Boundary markers: The skill explicitly requires the use of Zod schemas and factory guards (isXxxPayload) to validate all external data before processing.
  • Capability inventory: The skill executes shell commands (via pnpm and xl1 CLI), performs network operations (RPC and HTTP requests), and writes to the local file system.
  • Sanitization: Strict validation via Zod-first patterns is mandated in the dapp-checklist.md to ensure data integrity.
  • [EXTERNAL_DOWNLOADS]: The skill instructs the agent to download and install various packages and binaries.
  • Fetches official Node.js packages from the @xyo-network/* and @ariestools/* namespaces, which are associated with the skill's authoring organization.
  • Downloads the Chromium browser binary using the official playwright install command, a standard procedure for headless browser testing.
  • [COMMAND_EXECUTION]: The skill makes extensive use of CLI tools for environment setup and verification.
  • Uses the xl1 and xl1-wallet CLI tools for blockchain interactions.
  • Uses the security CLI (macOS) or secret-tool (Linux) to manage wallet passwords in the OS keychain, which is a recommended practice for secure automation.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 5, 2026, 09:38 PM