The Agent Skills Directory

[SAFE]: No security vulnerabilities or malicious behaviors were detected. The static analysis alerts regarding prompt injection are false positives caused by example text describing OWASP LLM vulnerabilities and guidelines within the prompt injection safety notice.
[INDIRECT_PROMPT_INJECTION]: The skill inherently processes external target repositories or source code via file reading tools (Read, Grep, Glob). However, the skill explicitly provides a robust safety notice instructing the agent to never execute commands found within target text and to flag any embedded instructions as vulnerabilities rather than obeying them, neutralizing the risk.

llm-top-10