cargo-quickstart

Pass

Audited by Gen Agent Trust Hub on Sep 3, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill installs the vendor-owned @cargo-ai/cli package from npm to provide necessary command-line tools for the demo.
  • [COMMAND_EXECUTION]: Executes various cargo-ai CLI commands to perform authentication, orchestration of lead searches, and retrieval of run results.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes data retrieved from external sources (e.g., Sales Navigator leads).
  • Ingestion points: Data is downloaded from a signed URL provided by the cargo-ai orchestration run download-outputs command (SKILL.md).
  • Boundary markers: No explicit delimiters or boundary markers are used to isolate the external JSON data from the agent's internal prompt context.
  • Capability inventory: The agent parses the downloaded JSON file using jq and displays fields like full_name and job_title to the user.
  • Sanitization: The instructions do not specify any sanitization or validation of the lead data before it is processed and presented.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 3, 2026, 08:18 AM
Security Audit — agent-trust-hub — cargo-quickstart