OUTCOMES
What this Prompt helps you deliver
- The integration contract is explicit
- Synchronization is retryable and auditable
BEFORE YOU START
Before you start
- Target API documentation and a test account are available
- Data ownership and synchronization direction are decided
What AI should inspect first
packages/appsapps/tabtin_djangodocs/agent/tooling-skills.mdREADY TO RUN
Complete Prompt
Copy the complete Prompt with its execution and acceptance boundaries.
Integrate an industry data source or business system into TabTin by defining the business contract and data responsibility before choosing API, Webhook, schedule, CLI, or connector. Read AGENTS.md, docs/agent/tooling-skills.md, packages/apps, and apps/tabtin_django. Collect documentation, test environment, authentication, samples, direction, frequency, scale, latency, failure ownership, sources of truth, cacheability, and sensitive fields.
Define object mapping, stable IDs, versions, deletion, and conflict policy. Design trigger, idempotency key, pagination, rate limiting, retry, resume, and dead letters. Store credentials safely with least privilege and no Token in source/logs. Register a discoverable App, connector, or CLI with clear Agent inputs/outputs. Record each sync’s scope, result, errors, replay data, and manual repair path.
Test initial full sync, increments, duplicates, out-of-order, limits, timeout, missing fields, expired permission, outage, and bounded volume. Retries must not duplicate business objects, cross-Organization data must not mix, and logs must not leak secrets. Report mapping, sequence, auth, errors/retries, metrics, tests, rollout, rollback, and external-admin actions without credentials.