Slickflow Documentation

Slickflow supports multi-agent orchestration using the ReAct (Reason → Act → Observe) loop. Each Agent node runs an autonomous loop:

1. Reason: Analyzes the task and available tools.
2. Act: Calls a registered tool (API, sub-agent, etc.).
3. Observe: Receives the tool result and decides the next step.

Agents use AgentConversationMemory to maintain per-session dialogue history across turns.

This test suite simulates a customer-AI interaction to verify the integration of CustomerService and MessageService with Supabase.

Workflow Steps:

1. Variable Initialization: Sets user_message (simulated customer input) and ai_response (simulated AI reply).
2. CustomerService Execution: Parses the user_message using regex to extract contact info (name, mobile, wechat, email). It then inserts or updates the biz_customer table in Supabase and writes the resulting customer_id to the process variables.
3. MessageService Execution: Saves the user_message, ai_response, and the customer_id (if available) to the biz_conversation table in Supabase.

Rules can be attached to gateway transitions in BPMN diagrams. When the engine reaches a split gateway, it automatically evaluates the rule expressions and follows the matching branch. Common patterns include:

• Amount threshold routing: e.g., amount > 10000 routes to a senior approval branch.
• Status-based branching: e.g., checking inventory levels.
• AI output routing: e.g., routing based on an LLM confidence score.

Slickflow supports traditional BPM patterns including Sequence, Split/Merge (AND/OR gateways), Sub-processes, and Multi-instance tasks.

Core human-task operations include:

• Start / Run: Launch or advance a process.
• Withdraw: Pull a task back from a user.
• SendBack: Return a task to a previous step.
• Resend / Reverse / Jump: Advanced exception handling and routing.

// Start a process instance
IWorkflowService wfService = new WorkflowService();
var startResult = wfService.CreateRunner("10", "Jack")
    .UseApp("DS-100", "Book-Order", "DS-100-LX")
    .UseProcess("PriceProcessCode")
    .Start();

// Run to next step
var runResult = wfService.CreateRunner("10", "Jack")
    .UseApp("DS-100", "Book-Order", "DS-100-LX")
    .UseProcess("PriceProcessCode")
    .NextStepInt("20", "Alice")
    .Run();

To run the Slickflow.Module.External.Tests (which simulates customer-AI chats and saves data to Supabase), you must configure your Supabase credentials. You can provide these via an appsettings.json file or through environment variables.

Option 1: appsettings.json

Ensure the following keys are present in your appsettings.json (and copied to the output directory):

• AiModelProvider:SupabaseProjectUrl
• AiModelProvider:SupabaseServiceRoleKey (use your service role key, or an anon key for testing)

Option 2: Environment Variables

Alternatively, set the following environment variables:

• SUPABASE_URL or SUPABASE_PROJECT_URL
• SUPABASE_SERVICE_ROLE_KEY or SUPABASE_ANON_KEY

{
  "AiModelProvider": {
    "SupabaseProjectUrl": "https://YOUR_PROJECT_REF.supabase.co",
    "SupabaseServiceRoleKey": "YOUR_SERVICE_ROLE_KEY"
  }
}

Slickflow uses SQL scripts located in the database/ directory to set up the schema and seed data.

Import Order:

1. Run the DDL script (wfdbtest2099_pgsql_schema.sql) to create tables, indexes, and sequences.
2. Run the DML script (wfdbtest2099_pgsql_data.sql) to insert initial seed and demo data.

Example using psql:

psql -h 127.0.0.1 -U postgres -d wfdbtest2099 -f database/wfdbtest2099_pgsql_schema.sql
psql -h 127.0.0.1 -U postgres -d wfdbtest2099 -f database/wfdbtest2099_pgsql_data.sql

Before importing the Slickflow database, ensure you have a target database created. You can use the createdb command-line tool. Adjust the host (-h), port (-p), user (-U), and database name as required by your environment.

createdb -h 127.0.0.1 -p 5432 -U postgres wfdbbpmn2

The Slickflow.Module.External.Tests requires specific tables in your Supabase project to function. You must ensure the following tables exist:

1. biz_customer: Stores customer contact information.
2. biz_conversation: Stores the chat history between the user and the AI.

You can create the biz_conversation table using the schema provided in Data/biz_conversation_supabase.sql within the project.

You can run the tests directly using the .NET CLI from the project directory, or by executing the built binary.

Note: If Supabase is not correctly configured, the program will exit with a short message and exit code 1.

cd source/core/Slickflow.Module.External.Tests
dotnet run