The Agent Skills Directory

[SAFE]: The skill serves as educational and reference material for developers using Claude Code. It contains no executable malicious logic or harmful instructions.
[SAFE]: The content proactively encourages security best practices, such as the use of the allowed-tools frontmatter field to enforce the principle of least privilege for bash commands (e.g., teaching users to use restricted patterns like Bash(git:*) instead of full access).
[SAFE]: Extensive documentation is provided on input validation and resource existence checks to prevent command failures and ensure robust command development.
[SAFE]: The YARA detection for destructive system commands in references/testing-strategies.md and references/marketplace-considerations.md is a false positive. The identified code snippets are examples of safe cleanup procedures for temporary directories (rm -rf $TEMP_DIR) or security test cases designed to verify that the agent correctly blocks unauthorized system access (e.g., testing that ls -la / is denied).

Command Development