onboarding-cro

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting data from external sources and user-supplied files.
  • Ingestion points: The agent is instructed to read .claude/product-marketing-context.md for context and use a Python script to process user-provided funnel.json data.
  • Capability inventory: The skill performs file system reads and executes a local Python script via the shell.
  • Boundary markers: The instructions lack specific boundary markers or directives to ignore instructions that might be embedded within the ingested data.
  • Sanitization: There is no evidence of content sanitization for the files processed by the agent or the script.
  • [COMMAND_EXECUTION]: The skill requires the execution of a bundled script (scripts/activation_funnel_analyzer.py) to perform its primary function. While the script logic is benign and handles local data, it involves the agent interacting with the shell.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 06:12 PM
Security Audit — agent-trust-hub — onboarding-cro