Security

The GGKP Knowledge Brokerage API is built on a Drupal-based REST architecture, chosen in part for its mature, enterprise-grade security framework. The security model operates across five distinct layers, each addressing a different class of risk. Together, they provide defence in depth: no single control is the sole barrier against unauthorised access or misuse.

1. Secure Authentication

Authentication is the outermost layer of access control. All API credentials are validated at the gateway before any service logic executes, and all communication is encrypted in transit. It is based on the issuing of an API KEY.

Important — API keys must be treated as secrets. Store them in environment variables or a secrets manager. Never commit keys to source code, client-side assets, or version control.

2. Role-Based Access Control

Once authenticated, the platform applies role-based access control (RBAC) to determine what an authenticated client is permitted to do. Access is enforced at the endpoint level and, where applicable, at the content type and field level.

Key properties of the RBAC model:

Each operation requires an explicit permission grant — there is no implicit access

Read, create, update, and delete permissions are independently assignable

Permission scope is fixed at key issuance and cannot be elevated at runtime

A partner with upload rights cannot delete products registered by another source

Permission reference:

Permission	Required For
`Upload Source`	`POST /search-federator/upload-source`
`Upload Products`	`POST /search-federator/upload-products`
`Domain Sources`	Domain-scoped source listing and management
`List Products`	`GET /search-federator/products/{source_id}`

Read operations — taxonomy retrieval, search queries, and PDF classification — require only a valid API key with no additional permission grants.

3. Data Exposure Control

Beyond controlling who can call the API, the platform controls what each call can return. The goal is to expose the minimum data necessary for each operation and eliminate unnecessary attack surface.

Controls in place:

Only explicitly required API endpoints are enabled — unused REST resources are disabled

Field-level restrictions prevent responses from returning data beyond what the operation requires

API responses are scoped to the minimum dataset needed for the requested function

Exposure configuration is reviewed whenever new endpoints or fields are introduced

4. CSRF and Request Protection

All write operations are protected against cross-site request forgery (CSRF). A valid API key alone is insufficient to execute a write from an unauthorised origin — the request must also carry a valid CSRF token.

Applies to: POST, PATCH, and DELETE operations.

CSRF token validation is enforced at the API gateway, before the request is routed to any service. This prevents malicious cross-site attacks from exploiting authenticated browser sessions.

5. Infrastructure-Level Protection

Security controls extend beyond the application layer to the infrastructure on which the API runs. These controls address network-level threats, volumetric abuse, and the operational visibility required to detect and respond to anomalous behaviour.

Control	Description
SSL/TLS	HTTPS enforced at the network boundary via SSL certificate
Web Application Firewall (WAF)	Filters malformed, malicious, and out-of-policy requests before they reach the application
Rate Limiting	Prevents API abuse and protects service availability across all clients
Monitoring & Logging	Comprehensive logging of API usage supports anomaly detection, incident response, and audit

Security Summary

Layer	Mechanism	What It Protects Against
Authentication	API key / OAuth 2.0 over HTTPS	Credential interception, unauthenticated access
RBAC	Per-endpoint permission grants	Unauthorised operations by authenticated clients
Data Exposure	Endpoint and field restrictions	Data leakage, over-broad API responses
CSRF Protection	Token validation on write operations	Cross-site request forgery
Infrastructure	WAF, rate limiting, TLS, logging	Network attacks, abuse, undetected anomalies

1. Secure Authentication#

2. Role-Based Access Control#

3. Data Exposure Control#

4. CSRF and Request Protection#