Your Name
83 lines
3.1 KiB
Gherkin
83 lines
3.1 KiB
Gherkin
Feature: Agent Management
|
|
As a CleverClaude user
|
|
I want to create, manage, and coordinate AI agents
|
|
So that I can build sophisticated AI workflows
|
|
|
|
Background:
|
|
Given CleverClaude is running
|
|
And I have agent management capabilities
|
|
|
|
@smoke
|
|
Scenario: Create a single agent
|
|
When I create a researcher agent named "research_agent_1"
|
|
Then the agent should be created successfully
|
|
And the agent should be in "active" status
|
|
And the agent should have researcher capabilities
|
|
|
|
Scenario: Create multiple agents with different types
|
|
When I create the following agents:
|
|
| type | name | capabilities |
|
|
| researcher| research_agent | research, analysis |
|
|
| coder | coding_agent | coding, debugging, testing |
|
|
| analyst | analyst_agent | data_analysis, visualization |
|
|
Then all agents should be created successfully
|
|
And each agent should have the correct type and capabilities
|
|
|
|
Scenario: Agent lifecycle management
|
|
Given I have created an agent named "test_agent"
|
|
When I pause the agent
|
|
Then the agent status should be "paused"
|
|
When I resume the agent
|
|
Then the agent status should be "active"
|
|
When I destroy the agent
|
|
Then the agent should no longer exist
|
|
|
|
Scenario: Agent task execution
|
|
Given I have a researcher agent
|
|
When I assign a research task to the agent:
|
|
"""
|
|
Research the latest developments in quantum computing
|
|
and provide a summary of the key breakthroughs
|
|
"""
|
|
Then the agent should accept the task
|
|
And the task should be executed within the timeout period
|
|
And the result should contain relevant research findings
|
|
|
|
Scenario: Agent health monitoring
|
|
Given I have multiple active agents
|
|
When I check agent health status
|
|
Then each agent should report health metrics
|
|
And unhealthy agents should be identified
|
|
And health metrics should include CPU, memory, and task count
|
|
|
|
Scenario: Agent capability discovery
|
|
Given I have agents with different capabilities
|
|
When I query for agents with "data_analysis" capability
|
|
Then only agents with that capability should be returned
|
|
And the results should include agent metadata
|
|
|
|
@wip
|
|
Scenario: Agent coordination
|
|
Given I have multiple agents of different types
|
|
When I create a coordination task requiring multiple agent types
|
|
Then the agents should coordinate automatically
|
|
And the task should be distributed appropriately
|
|
And the results should be aggregated correctly
|
|
|
|
Scenario Outline: Create agents with various configurations
|
|
When I create a <type> agent with <timeout> second timeout
|
|
Then the agent should be created with the specified timeout
|
|
And the agent should respect timeout limits during task execution
|
|
|
|
Examples:
|
|
| type | timeout |
|
|
| researcher | 30 |
|
|
| coder | 60 |
|
|
| analyst | 45 |
|
|
|
|
@hypothesis
|
|
Scenario: Stress test agent creation
|
|
When I create many agents rapidly
|
|
Then all agents should be created successfully
|
|
And system performance should remain stable
|
|
And no resource leaks should occur |