The Agent Skills Directory

[COMMAND_EXECUTION]: The skill mandates a 'Code Reuse Analysis' using grep and codebase_search to explore local files for existing logic. These operations are restricted to the local filesystem and are intended for project efficiency.
[EXTERNAL_DOWNLOADS]: The skill references Docker images and CI/CD templates from trusted internal Kavak sources, including a private Amazon ECR registry and gitlab.com/kavak-it/ci-jobs.
[PROMPT_INJECTION]: While the skill's reliance on searching local code for logic reuse creates a potential surface for indirect prompt injection from comments within the codebase, this is a standard development workflow and the skill itself contains no malicious instructions.
[SAFE]: The skill correctly uses placeholders for sensitive credentials (e.g., my-secret) and adheres to secure coding practices like using parameterized queries to prevent injection vulnerabilities.

backend-development