The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains only instructional content for shell scripting. No injection or override patterns were detected.
[DATA_EXFILTRATION]: No patterns for unauthorized data access or external transmission were found. The skill promotes secrets detection using tools like Gitleaks.
[REMOTE_CODE_EXECUTION]: The instructions explicitly warn against the use of eval on untrusted input and advocate for checksum verification when sourcing external scripts.
[EXTERNAL_DOWNLOADS]: Mentions standard community package managers such as basher and bpkg, and established testing/linting tools like ShellCheck and Bats-core as part of a secure development lifecycle.
[SAFE]: The content is dedicated to hardening shell scripts and implementing safety patterns such as strict mode (set -Eeuo pipefail), proper quoting, and restrictive file permissions.

bash-pro