Security audit
Security checks for vulnerabilities and agentic risk
This skill compares two chosen files and formats the differences, with no evidence of hidden persistence, network activity, credential handling, or destructive behavior.
Install only if you are comfortable letting the agent compare file paths you explicitly provide. Avoid asking it to diff secrets, credentials, private keys, or other sensitive files unless you intend their contents to appear in the response.
The skill instructs use of a shell command (diff / diff -u) but does not declare any permissions or allowed-tools scope. That creates an undeclared capability boundary: an agent may invoke shell access for arbitrary file comparisons without explicit restriction, increasing the chance of unintended access to sensitive local paths or misuse of shell execution in a broader runtime.
The trigger language is broad and includes generic phrases like diff and show me the differences, which could match normal conversation and cause the skill to activate unexpectedly. In a skill that can drive shell-based file access, accidental activation raises the risk of comparing unintended files or exposing sensitive file contents through overbroad routing.
subprocess module calls execute external commands. Without careful input validation, this enables command injection.
def run_diff(source, target):
"""Run diff command and return unified output."""
try:
result = subprocess.run(
["diff", "-u", source, target],
capture_output=True,
text=True
No suspicious patterns detected.