agent-onboarding

Fail

Audited by Gen Agent Trust Hub on Oct 29, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: Provides instructions to download and execute the Bright Data CLI installation script from the vendor's official domain (https://cli.brightdata.com/install.sh).
  • [EXTERNAL_DOWNLOADS]: Facilitates the installation of the Bright Data CLI and SDKs via NPM and PyPI, targeting official vendor-maintained packages such as @brightdata/cli and brightdata-sdk.
  • [INDIRECT_PROMPT_INJECTION]: The skill establishes a workflow for ingesting untrusted data from the open web, creating an attack surface for indirect prompt injection. • Ingestion points: Content from external URLs retrieved via scraping and search tools like bdata search and bdata scrape. • Boundary markers: No specific delimiters or warnings against embedded instructions are provided in the onboarding flow. • Capability inventory: The skill enables shell command execution through the Bright Data CLI and network operations through official SDKs. • Sanitization: No sanitization or validation of the fetched web content is performed at this stage.
Recommendations
  • HIGH: Downloads and executes remote code from: https://cli.brightdata.com/install.sh - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Oct 29, 2026, 04:30 PM