The Agent Skills Directory

[SAFE]: The skill serves as a technical reference for building robust backend systems. It outlines best practices for durability, idempotency, and concurrency control using established technologies like PostgreSQL, Redis, and Convex.
[INDIRECT_PROMPT_INJECTION]: The architectural patterns described (such as file imports in ASYNC-APIS.md and event-driven workflows in DURABLE-WORKFLOWS.md) create surfaces for processing untrusted data.
Ingestion points: ASYNC-APIS.md (POST endpoints for file imports and media processing), DURABLE-WORKFLOWS.md (waiting for external events or human replies).
Boundary markers: No specific delimiters or safety instructions are defined in the provided architectural guidelines.
Capability inventory: The guide covers database interactions (SQL), storage uploads (Supabase/Convex), and external API coordination (webhooks).
Sanitization: The instructions focus on system durability and atomicity rather than data sanitization for AI processing.
[COMMAND_EXECUTION]: While the skill contains SQL code snippets and pseudocode in WORKERS.md and ASYNC-APIS.md, these are provided as implementation examples for the developer and do not involve the agent executing arbitrary shell commands or untrusted input.

background-jobs