Security audit
Security checks for vulnerabilities and agentic risk
The skill is a disclosed GOG cleanup helper that can email a report and create Apple Reminders, with no hidden payloads found.
Before installing, confirm you want an agent to use your configured himalaya email account and Apple Reminders access for this workflow. Use --dry-run first, check the recipient address, and use --no-email or --no-reminders when you only want analysis.
The skill describes and encourages shell-capable actions (python3, himalaya, and remindctl) but does not declare any tool scope or permissions boundaries. This creates an authorization/expectation gap where an agent may execute side-effecting commands without explicit least-privilege constraints, increasing the chance of unintended email sending or reminder creation.
The 'Use when' wording is broad enough that an agent could invoke this skill for loosely related cleanup or notification requests, even though it performs external side effects like sending email and creating reminders. Over-broad triggers increase the risk of the skill running in contexts the user did not specifically authorize, leading to unwanted messaging or task creation.
subprocess module calls execute external commands. Without careful input validation, this enables command injection.
"-H", f"Subject:{subject}",
body,
]
result = subprocess.run(cmd, capture_output=True, text=True)
if result.returncode != 0:
print(f"Error sending email: {result.stderr}", file=sys.stderr)
sys.exit(1)
subprocess module calls execute external commands. Without careful input validation, this enables command injection.
"-H", f"Subject:{subject}",
body,
]
result = subprocess.run(cmd, capture_output=True, text=True)
if result.returncode != 0:
print(f"Error sending email: {result.stderr}", file=sys.stderr)
sys.exit(1)
No suspicious patterns detected.