| name | exint - External Integration |
| description | Design how something connects to external systems, tools, or services. Identifies touchpoints, maps data flows, defines interfaces, assesses dependencies, plans for failures, and produces an integration checklist. |
| output | {"format":"prose"} |
External Integration
Input: $ARGUMENTS
Step 1: Identify External Touchpoints
List every point where the system touches something external.
SYSTEM: [what is being integrated]
INTEGRATION GOAL: [why external connections are needed]
EXTERNAL TOUCHPOINTS:
1. [system/service] — Direction: [IN / OUT / BOTH] — Purpose: [why]
2. [system/service] — Direction: [IN / OUT / BOTH] — Purpose: [why]
...
TOUCHPOINT CATEGORIES:
- Data sources: [list]
- Data sinks: [list]
- Services consumed: [list]
- Services provided: [list]
- User-facing integrations: [list]
Step 2: Map Data Flows
Trace what data moves where, in what format, and how often.
DATA FLOWS:
[Source] →(format, frequency)→ [Destination]
1. [external] →([format], [frequency])→ [internal]
Data: [what's transferred]
Volume: [expected size/rate]
2. [internal] →([format], [frequency])→ [external]
Data: [what's transferred]
Volume: [expected size/rate]
TRANSFORMATIONS NEEDED:
- [flow]: [input format] must become [output format] via [method]
DATA OWNERSHIP: [who owns each data flow?]
Step 3: Define Interfaces
Specify the contracts between systems.
INTERFACES:
1. [interface name]
- Protocol: [REST / GraphQL / webhook / file / etc.]
- Authentication: [method]
- Contract: [what's guaranteed]
- Versioning: [how changes are handled]
2. [interface name]
- Protocol: [type]
- Authentication: [method]
- Contract: [what's guaranteed]
- Versioning: [how changes are handled]
STANDARDS TO FOLLOW: [relevant standards or conventions]
DOCUMENTATION REQUIRED: [what needs to be documented for each interface]
Step 4: Assess Dependencies
Evaluate the risk profile of each external dependency.
DEPENDENCY ASSESSMENT:
| External System | Criticality | Reliability | Control | Risk |
|----------------|-------------|-------------|---------|------|
| [system 1] | [HIGH/MED/LOW] | [HIGH/MED/LOW] | [OWN/SHARED/NONE] | [score] |
| [system 2] | [HIGH/MED/LOW] | [HIGH/MED/LOW] | [OWN/SHARED/NONE] | [score] |
SINGLE POINTS OF FAILURE:
- [dependency]: if this fails, [consequence]
VENDOR LOCK-IN RISKS:
- [dependency]: switching cost is [HIGH/MED/LOW] because [reason]
Step 5: Plan for External Failures
Design resilience for when external systems fail.
FAILURE SCENARIOS:
1. [external system] goes down
- Detection: [how we know]
- Impact: [what breaks]
- Mitigation: [fallback / cache / queue / degrade gracefully]
- Recovery: [how to resume when it's back]
2. [external system] returns bad data
- Detection: [validation approach]
- Impact: [what breaks]
- Mitigation: [reject / sanitize / alert]
CIRCUIT BREAKERS: [where to stop calling failing externals]
RETRY STRATEGY: [when and how to retry]
TIMEOUT POLICY: [how long to wait before giving up]
Step 6: Integration Checklist
INTEGRATION CHECKLIST:
PRE-INTEGRATION:
☐ All interfaces documented
☐ Authentication credentials secured
☐ Data formats validated
☐ Error handling defined
☐ Rate limits understood
IMPLEMENTATION:
☐ Each touchpoint implemented and tested
☐ Data transformations verified
☐ Failure modes handled
☐ Logging and monitoring in place
☐ Circuit breakers configured
POST-INTEGRATION:
☐ End-to-end flow tested
☐ Performance validated under load
☐ Failover tested
☐ Documentation updated
☐ Runbook created for operations
PRIORITY ORDER:
1. [first integration to build] — Reason: [why first]
2. [second integration] — Reason: [why second]
Integration
Use with:
/ecomp -> Assess the full ecosystem including external connections/gapf -> Find missing integrations/de -> Plan the integration project execution
