The Agent Skills Directory

[COMMAND_EXECUTION]: The skill orchestrates a wide array of shell commands to manage environment checks, dependency auditing (via pip-audit), and testing frameworks.
[DYNAMIC_EXECUTION]: Employs python manage.py shell and heredoc strings to execute runtime Python code for validating configuration settings, cache reachability, and database integrity.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: The skill reads and processes source code files, Git diffs, and tool-generated security reports (e.g., bandit-report.json).
Boundary markers: No specific boundary markers or instructions are provided to the agent to ignore potential prompts embedded within the analyzed project data.
Capability inventory: The skill possesses the ability to execute shell commands, perform network operations via security scanners, and write files to the local system.
Sanitization: The skill lacks explicit sanitization steps for data ingested from the repository environment before it is processed by the agent.

django-verification