The Agent Skills Directory

[SAFE]: The skill provides documentation and code examples for implementing common architectural patterns such as task queues, idempotency, and job state machines.
[INDIRECT_PROMPT_INJECTION]: The skill documents patterns for processing external data (e.g., webhooks, API requests) in background workers. While this defines an ingestion surface for untrusted data, the code examples follow standard engineering practices (such as using Pydantic and repository patterns) and the risk is inherent to the primary purpose of building background processing systems.
Ingestion points: ExportRequest model in SKILL.md, webhook_id and payload in references/details.md.
Boundary markers: None explicitly implemented in the template code.
Capability inventory: Database interactions via jobs_repo, network connectivity via Celery brokers, and interaction with external service clients (email, payments).
Sanitization: Uses Pydantic (model_dump) and type hinting for basic data validation.

python-background-jobs