The Agent Skills Directory

[Informational Vulnerability Database]: The skill serves as a static reference for identifying and mitigating LLM security risks. It does not include any scripts, executable files, or automation that would perform actions on the system or access external resources.
[Reference Security Examples]: Several files contain illustrative examples of malicious prompts, such as those found in the prompt injection and vector weaknesses references. These examples are documented as attack scenarios for identification and testing purposes and are not active instructions for the agent to execute.
[No External Dependencies]: The skill does not define or require any external Python or Node.js packages, nor does it attempt to download or execute remote scripts at runtime.
[Metadata Consistency]: The skill's description and contents are consistent with its stated purpose of providing a vulnerability knowledge base based on established OWASP security standards.

owasp-llm