client-onboarding

Pass

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: Indirect prompt injection surface via external data ingestion.
  • Ingestion points: SKILL.md (Steps 1, 2, 3, 7) fetches content from company websites, LinkedIn profiles, and industry news.
  • Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the playbook.
  • Capability inventory: SKILL.md instructions involve directory creation (clients/) and file writing for intelligence packages and strategies.
  • Sanitization: No content sanitization before processing.
  • [COMMAND_EXECUTION]: The playbook in SKILL.md directs the agent to create a directory hierarchy and write multiple markdown files locally under the clients/ folder.
  • [EXTERNAL_DOWNLOADS]: SKILL.md utilizes web scraping tools and fetching capabilities to gather intelligence from various external domains, including LinkedIn and ad platforms.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 29, 2026, 05:18 PM