The Agent Skills Directory

[SAFE]: The skill contains no executable code or scripts; it is comprised entirely of markdown documentation and reference guides for security practitioners.
[INDIRECT_PROMPT_INJECTION]: The skill proactively identifies risks associated with AI coding agents in CI/CD pipelines. It identifies PR comments, issue bodies, and fork content as untrusted ingestion points and provides appropriate remediation guidance, such as requiring human review and sandboxing agent actions.
[COMMAND_EXECUTION]: The documentation provides examples of shell commands and pipeline configurations for the purpose of demonstrating security gates. It specifically identifies and warns against dangerous execution patterns, such as piping unverified remote content into a shell.
[EXTERNAL_DOWNLOADS]: The skill references standard security tools and GitHub Actions from well-known and trusted providers (e.g., AWS, Semgrep, Aqua Security) as part of its architectural recommendations.

devsecops