SkillMachine
All skills

API Integration Mapper

Report

Maps the integration surface between two systems: field mapping, auth plan, and a resilience score

Live output preview

Input Format: Input FormatOutputWatch the Output: Watch the Output

A plan is required to view this content

Choose a plan to access input format, sample outputs, and live previews.

View Plans →

About the skill

What it does

Turns the integration surface between two systems into a concrete plan an integration engineer can drop into a sprint. It binds to five canonical frameworks — not intuition.

Integration style decision. Based on the trigger's nature it picks and justifies synchronous REST/RPC (low latency, tight coupling), webhook→queue async (loose coupling, durable), or batch/bulk (high volume, rate-limit friendly). When 3+ systems share the same data it recommends a Canonical Data Model (Hohpe & Woolf), which cuts translators from N² to ~2N (12 instead of 30 for 6 systems).

Field mapping (Message Translator). Each A-field → B-field row: transform type (1:1 / format / merge / split / constant / lookup), required-ness, data type, example. An unmappable required B-field is a critical finding (422 on B). An unmappable A-field is an information-loss warning.

Auth plan. Decision tree: user delegation → Authorization Code + PKCE; system-to-system → Client Credentials; simple machine → API key + HMAC signature; regulated → mTLS. Token TTL, refresh, secret storage (env/vault — never in code), and scope minimization are resolved. For webhooks, HMAC signature + replay protection (timestamp+nonce) is a mandatory check.

Resilience score. A 6-criterion, 0-100 weighted rubric: idempotency (25), retry+backoff+jitter (20), error classification (15), rate-limit awareness (15), outbox/DLQ (15), observability (10). Score → verdict: 0-39 Fragile (demo), 40-69 Borderline (pilot-only), 70-100 Production-grade.

Breaking-change scan. Whether A's schema change is backward-compatible (add field = MINOR) or breaking (remove field = MAJOR), and whether B is a tolerant reader — flagging which change breaks which consumer.

When to use it

  • "How do I connect X to Y", "I'm setting up a webhook integration to this API".
  • A robustness audit of an existing integration (idempotent? retries? which auth? which rate-limit wall?).
  • You have an OpenAPI/Postman collection, endpoint list, or webhook payload and want a field-mapping + auth + resilience plan.
  • Moving n8n/Zapier/Make + custom REST/webhook integrations from "demo" to "production".

Method / frameworks

  1. Enterprise Integration Patterns (Hohpe & Woolf, 65 patterns).
  2. Integration Style selection — RPC vs Messaging vs File vs Shared-DB.
  3. OAuth 2.0 / OIDC taxonomy — RFC 6749/6750, PKCE 7636, JWT 7519, RFC 8252.
  4. Reliability/Resilience set — idempotency, retry+exponential backoff+jitter, circuit breaker, outbox/DLQ.
  5. API Evolution / Contract Compatibility — semantic versioning + tolerant-reader.

How do I use this skill?

You don't "run" a skill — after installing it you just tell the agent your task (e.g. ask for the relevant job), and the skill kicks in by itself when its description matches.

Upload the api-entegrasyon-haritalayici.zip you downloaded as-is — no packaging needed, the format is already correct (folder at root).

  1. Open Settings → Customize → Skills
  2. Upload → select the api-entegrasyon-haritalayici.zip you downloaded
  3. Claude reads SKILL.md; the name + description appear. Ready ✅

Scripts run in Anthropic's code-execution environment (sandbox) — not on your machine.