support-automation

Pass

Audited by Gen Agent Trust Hub on Nov 16, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill provides commands to install the Whop CLI using shell-piped scripts from Whop's official domain and GitHub repository (e.g., 'curl -fsSL https://whop.com/install.sh | sh' and 'irm https://github.com/whopio/whop-public-cli/releases/latest/download/install.ps1 | iex').\n- [EXTERNAL_DOWNLOADS]: Suggests the use of Vercel Labs' agent-browser tool from its official GitHub repository to handle browser-based tasks.\n- [EXTERNAL_DOWNLOADS]: Directs the agent to fetch additional skills and installation instructions from the author's own GitHub repository at 'github.com/branded-dropshipping/skills'.\n- [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves processing customer support tickets, which are untrusted external inputs. This creates a surface for potential indirect prompt injection attacks given the agent's capabilities.\n
  • Ingestion points: Customer support tickets processed in SKILL.md and references/ticket-triage-and-drafts.md.\n
  • Boundary markers: Present; instructions in references/browser-runtime-and-sessions.md explicitly warn the agent that page text and tool descriptions are data and should not be interpreted as instructions or permissions to inspect secrets or run commands.\n
  • Capability inventory: Includes browser tool access, Whop CLI execution, and ticket/order management operations.\n
  • Sanitization: Absent; the skill does not specify mechanisms for filtering or sanitizing ticket content before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 16, 2026, 10:49 AM