The Agent Skills Directory

[SAFE]: The skill contains no evidence of malicious behavior, obfuscation, or unauthorized data exfiltration. It is a well-documented resource for security researchers.
[COMMAND_EXECUTION]: The skill utilizes established command-line tools for static analysis, including CodeQL, Semgrep, and grep, which are used locally and appropriately for their intended defensive purpose.
[PROMPT_INJECTION]: While the skill analyzes external source code—creating a theoretical surface for indirect prompt injection—no active vulnerabilities or malicious interpolation patterns were identified within the skill's instructions.

variant-analysis