sales-clean-email

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an Indirect Prompt Injection surface by ingesting and modifying local reference files without using security delimiters or sanitization. \n
  • Ingestion points: The skill reads references/learnings.md and references/platform-guide.md to gather context and platform knowledge. \n
  • Boundary markers: Absent. No delimiters are used to separate untrusted file content from instructions. \n
  • Capability inventory: The skill can read both reference files and append new information to references/learnings.md. \n
  • Sanitization: Absent. The skill does not validate or filter content before processing or appending. \n- [COMMAND_EXECUTION]: The skill provides instructions for the user or agent to execute shell commands to download and install additional related skills. \n
  • Evidence: The 'Related skills' section suggests multiple commands using the pattern npx skills add sales-skills/sales --skill [skill-name] -a claude-code for services like SaneBox, Inbox Zero, and Superhuman.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 01:00 AM
Security Audit — agent-trust-hub — sales-clean-email