The Agent Skills Directory

[SAFE]: The skill focuses on providing best practices and code snippets for browser automation (Playwright), media processing (ffmpeg, ImageMagick, Sharp), and shell scripting. No malicious behavior or insecure patterns were detected.
[COMMAND_EXECUTION]: The script scripts/check_toolchain.sh performs routine environment checks using command -v and directory existence tests to ensure necessary tools like Node.js and browser binaries are present. These operations are benign and typical for development environments.
[DATA_EXFILTRATION]: The skill explicitly advises users to manage sensitive data like session cookies using environment variables and .env files, which are standard security practices to prevent credential exposure in source control.
[PROMPT_INJECTION]: The use of $ARGUMENTS in SKILL.md is limited to providing the agent with a task boundary or scope (e.g., a repository path or component name). It does not involve executing user-supplied input as shell commands or system instructions.
[EXTERNAL_DOWNLOADS]: The skill references standard, well-known libraries and tools (Playwright, Sharp, dotenv, ffmpeg). It does not include commands to download or execute code from untrusted remote sources.

automation-builder