Skip to main content
TealTiger modules don’t operate in isolation. In production, they compose into governance pipelines where each module handles a specific checkpoint — input scanning, policy evaluation, tool authorization, memory governance, secret detection, and audit logging. These pages walk through complete orchestration flows for common use cases. Each flow shows exactly which modules fire, in what order, and what they enforce at every step.

​
Use cases

Customer Support Agent

Tool authorization, human-in-the-loop approval gates, cost governance, and full audit trails for support automation.

Code Assistant

Secret detection in generated code, memory scope governance, and model allowlisting for AI coding tools.

RAG Application

Input/output governance, memory read classification, and PII redaction for retrieval-augmented generation.

Multi-Agent System

Delegation governance, privilege boundaries, and correlation ID tracing across cooperating agents.

Healthcare Agent (HIPAA)

PHI/PII governance, strict memory classification, HIPAA-compliant audit evidence, and SARIF export.

Financial Advisor Agent

Approval gates for high-value trades, circuit breakers on external APIs, and regulatory audit trails.

​
How to read these flows

Each orchestration page includes:
  • Step-by-step flow — Every governance checkpoint in order, with the module responsible
  • Flow diagram — Visual pipeline showing the full request lifecycle
  • Code example — Complete TypeScript implementation you can adapt
  • Governance points table — Summary of which modules fire and what they enforce
  • Related docs — Links to API reference and concept pages for deeper dives
These flows use TealTiger v1.2 modules (TealEngine, TealGuard, TealSecrets, TealRegistry, TealMemory, TealReliability, TealAudit). If you’re on v1.1.x, see the migration guide for module availability.

​
Composability principle

TealTiger’s governance is deterministic and composable. Every module produces a typed ModuleResult that feeds into TealEngine’s merge pipeline. The engine applies a “most restrictive action wins” strategy — if any module says DENY, the request is denied regardless of what other modules allow. This means you can add or remove modules from any flow without breaking the pipeline. Start with TealGuard for input scanning, add TealSecrets when you need output scanning, layer in TealRegistry when you need tool allowlisting — each module slots in independently.
All orchestration flows produce a correlation ID that traces the full decision chain. Use this ID to query TealAudit for the complete evidence envelope.