The Agent Skills Directory

[PROMPT_INJECTION]: The skill documentation includes examples of prompt injection payloads, such as "Ignore previous instructions". These are listed explicitly as threat patterns for the scanner to detect in other files and do not function as instructions to the model to bypass its own safety guidelines.
[NO_CODE]: The skill consists entirely of markdown documentation and YAML metadata. It lacks any functional implementation code (Python, Shell, or JavaScript) to carry out the described security scans, functioning instead as a set of rules for the agent.
[SAFE]: The skill's stated purpose and content are entirely defensive. It provides a framework for identifying common attack vectors like shell injection and secret exfiltration without introducing any malicious behavior itself.

aeon-skill-security-scan