The Agent Skills Directory

[SAFE]: The skill defines a standard set of professional development guidelines for Django projects, including templates for models, views, and testing. It explicitly instructs the agent to follow security best practices, such as avoiding raw SQL, using environment variables for secrets, and implementing proper permissions.
[INDIRECT_PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection because it instructs the agent to ingest and act upon external data such as user requirements, existing codebases, and tool outputs to generate implementation code.
Ingestion points: Project requirements, model definitions, existing settings.py files, and CLI output from manage.py commands.
Boundary markers: The skill does not provide explicit boundary markers or instructions for the agent to distinguish between its internal logic and untrusted data provided by the user or the environment.
Capability inventory: The agent is empowered to perform file system writes (creating/modifying source code), execute shell commands (database migrations via manage.py), and configure authentication mechanisms (JWT, session auth).
Sanitization: While the skill provides templates for application-level data sanitization (via DRF serializers), it lacks sanitization or validation protocols for the input data processed by the agent during the task execution itself.

django-expert