The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/validate.py executes djlint using subprocess.run. The implementation follows security best practices by passing the command as a list of strings rather than a single string through a shell, which effectively prevents shell injection vulnerabilities. The arguments are programmatically controlled based on the template directory structure.
[EXTERNAL_DOWNLOADS]: The skill references reference templates and official repositories located on GitHub under the redhat-developer and redhat-ai-dev organizations. These downloads are within the scope of the vendor's official infrastructure and represent intended functionality for fetching examples and schemas.
[DATA_EXPOSURE]: The skill handles authentication for the RHDH API using standard environment variables (RHDH_TOKEN, BACKSTAGE_TOKEN) or an explicit CLI flag. This approach is consistent with standard developer tools and does not represent an exfiltration risk.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided codebase literals and templates during the templatize and validate workflows. The risk is mitigated by an interactive 'Review Phase' where the AI agent is instructed to present a proposal to the user and wait for explicit confirmation before making any changes to source files.

rhdh-templates