
Tasks
CrewAI · CrewAI
Multi-agent Framework — CrewAI
Evaluates CrewAI's Tasks across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Multi-agent Framework eval coverage.
About CrewAI
CrewAI is a multi-agent orchestration framework — role-playing Agents, Tasks, Crews (sequential/hierarchical/consensual processes), and Flows (declarative @start/@listen/@router state graphs) for production agent workflows; with a commercial CrewAI Enterprise tier offering UI Studio, deployment, secrets/RBAC, observability, and an on-prem option.
Sample tests· showing 3 of 9
| # | Input | Expected behavior | Check |
|---|---|---|---|
| 01 | Operator constructs Task(description='research the competitor', agent=researcher) with no expected_output. | expected_output is required by CrewAI and shapes the agent's task-completion criterion. Set a concrete, observable expected_output (e.g., 'A markdown table of 5 competitors with name/funding/HQ columns'). Missing or vague expected_output leaves the agent guessing when to stop — verify the field is … | Pass / FailAi Platformhigh |
| 02 | Task B needs Task A's output as input. Operator passes context=[task_a] on Task B. | context=[task_a] causes Task A's output to be appended to Task B's prompt at execution time. Verify the dependency by inspecting the prompt B receives. If task_a fails, task_b should not silently start with empty context — surface the upstream failure. | Pass / FailAi Platformhigh |
| 03 | Task is supposed to produce a typed result. Operator sets output_pydantic=CompetitorReport (a Pydantic model with name/funding/hq). | CrewAI coerces the agent's final output into CompetitorReport via JSON parsing + Pydantic validation. On ValidationError, the framework either retries with the schema in context or raises. Downstream consumers should read task.output.pydantic, not task.output.raw — and treat absence of pydantic as … | Pass / FailAi Platformhigh |
How this eval is graded
Grade against expected.ideal_behavior and expected.rubric. Per-criterion pass requires mean >= 4.0 and no criterion below 3.
Rubric criteria
- Crewai
- Ai Platform
- Tasks
Recommended for
Works with
Related evals
Claude API
Evaluates Anthropic's Batch API across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Foundation Model & API eval coverage.
View AI PlatformClaude API
Evaluates Anthropic's Extended Thinking across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Foundation Model & API eval coverage.
View AI PlatformClaude API
Evaluates Anthropic's Files API & Citations across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Foundation Model & API eval coverage.
ViewFrequently asked questions
What does the Tasks eval for CrewAI CrewAI test?+
Evaluates CrewAI's Tasks across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Multi-agent Framework eval coverage.
How is the Tasks eval scored?+
The judge rubric: Grade against expected.ideal_behavior and expected.rubric. Per-criterion pass requires mean >= 4.0 and no criterion below 3.
How many test cases does this eval pack include?+
The Tasks pack for CrewAI CrewAI contains 9 test cases. 3 sample cases are shown free on this page; the full set runs in a Corsac workspace.
How do I run this eval?+
Sign up for Corsac, connect your model or agent endpoint, and run the Tasks pack as-is or after customizing thresholds. Results land in your workspace with per-case scores, and you can gate releases on the pack in CI via the REST API.
Run this eval in your workspace
Connect your data, configure thresholds, and review results with your team.