Agent skill · software engineering · jeremylongshore
attio-reference-architecture
Design a production Attio integration architecture with tenant-bound credentials, schema discovery, queued writes, verified webhooks, idempotency, reconciliation, and rollback ownership. Use when planning or reviewing a multi-workspace Attio service. Trigger with "Attio architecture", "design Attio integration", or "Attio system design".
Why this skill is useful
Adds a complete production reference architecture for Attio CRM integrations, including executable code examples and specific API usage patterns not commonly found in public documentation.
What it needs
Requires bull, redis installed locally. Requires attio account access. About 2k tokens when loaded. Last updated 2026-08-07. 2,604 stars on the source repository.
What this skill does
Attio Integration Reference Architecture Overview This skill produces a repo-grounded architecture for reliable Attio reads, writes, and events. It treats tenant isolation, duplicate delivery, schema evolution, and reconciliation as first-class controls. Prerequisites Workspace and tenancy model Required objects, lists, attributes, records, entries, and events Availability, freshness, recovery, and retention objectives Owners for credentials, queues, data mapping, and incident response Tool Discipline Use Read, Glob, and Grep to map components, trust boundaries, state, and existing clients. Use WebFetch only for current official Attio contracts. Use Write or Edit after the architecture, ownership, and approval boundaries are agreed. Current Contract Use OAuth for multi-workspace applications and a workspace key only for a controlled single-workspace integration. Discover object and attribute identifiers rather than hard-coding display labels. Pagination differs by endpoint, and record or entry queries can carry score-based rate cost. Webhooks are at-least-once signals; durable processing needs signature verification, idempotency, a queue, and reconciliation. Authentication Place a tenant credential resolver behind a server-side interface. Bind each encrypted token to one workspace identity and exact scopes; never accept a workspace identifier from a caller without authorization. Instructions 1. Map ingress, egress, data stores, trust boundaries, tenant context, and failure domains from the repository. 2. Define a credential resolver and auditable endpoint-to-scope map. 3. Add a bounded schema registry or cache with explicit refresh and invalidation. 4. Route outbound mutations through idempotent jobs with separate read and write governors. 5. Terminate webhooks at an HTTPS receiver that verifies the raw body, deduplicates by idempotency key, acknowledges quickly, and queues work. 6. …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills jeremylongshore/attio-reference-architecture