marketing-report-expert

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill features an indirect prompt injection surface as it ingests untrusted data from external websites (via browser_subagent) and user-provided files to determine report styling and content.
  • Ingestion points: External URLs provided by users and local text files (e.g., data.txt).
  • Boundary markers: No specific delimiters or instructions to ignore embedded commands are present in the prompt templates.
  • Capability inventory: The skill can Read, Write, and RunCommand (specifically minimax-pdf).
  • Sanitization: No explicit sanitization or validation of the extracted style tokens or file content is described before they are passed to other tools.
  • [COMMAND_EXECUTION]: The skill uses the RunCommand tool to execute minimax-pdf. It follows a 'One-Prompt Automation' rule that chains crawling, extraction, and rendering without pausing for user feedback, which reduces the opportunity for human review of the parameters being passed to the command-line tool.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 02:17 PM
Security Audit — agent-trust-hub — marketing-report-expert