The Agent Skills Directory

[Command Execution] (HIGH): The skill uses a shell command git diff $1 --stat where $1 is a user-provided branch name. This is vulnerable to command injection if the branch name contains shell metacharacters (e.g., ;, &, |).
[Remote Code Execution] (HIGH): The skill instructs the agent to run npm test after modifying code. If the repository or branch being analyzed contains a malicious test script in package.json, the agent will execute arbitrary code with the user's privileges.
[Indirect Prompt Injection] (HIGH): The skill processes external content (git diffs) and has the capability to write files and execute commands (npm test). An attacker could embed malicious instructions or code in a branch that, when processed by this skill, leads to unauthorized code execution or data modification.
[Sanitization] (LOW): There is no evidence of input validation or sanitization for the branch name or the code being processed.

slop