Leadership Calibration powered by CrewAI

3_crew/community_contributions/leadership_calibration/README.md

@@ -0,0 +1,123 @@
+## Leadership Calibration Crew
+
+Welcome to the **Leadership Calibration** project, powered by [crewAI](https://crewai.com).
+
+This crew models a structured debate between two senior leadership personas:
+
+- **Senior Architect** – focuses on technical design, scalability, and system quality.
+- **Engineering Manager** – focuses on people, delivery, and business impact.
+
+Given a topic, the crew runs a multi-step debate and produces:
+
+- **Architect position** → `architect.md`
+- **Engineering manager position** → `manager.md`
+- **Final synthesized conclusion** → `conclusion.md`
+
+You can run the crew through an interactive Gradio UI.
+
+---
+
+## Installation
+
+- **Python**: requires `>=3.10,<3.14`
+- **Dependencies**: managed with [UV](https://docs.astral.sh/uv/) and `pyproject.toml`
+
+Install `uv` (once, globally):
+
+```bash
+pip install uv
+```
+
+From the project root (`3_crew/leadership_calibration`), install dependencies:
+
+```bash
+uv sync
+```
+
+Or, if you prefer the crewAI CLI workflow:
+
+```bash
+crewai install
+```
+
+Make sure your `OPENAI_API_KEY` (or other LLM provider keys) is available in the environment
+or defined in a local `.env` file.
+
+---
+
+## Project Structure
+
+Key components of the crew live under `src/leadership_calibration`:
+
+- **Agents configuration**: `config/agents.yaml`  
+  - `engineering_manager_agent` – senior engineering manager & technical leader  
+  - `senior_architect_agent` – senior software engineer & architect
+
+- **Tasks configuration**: `config/tasks.yaml`  
+  - `architect_position_statement` → writes `output/architect.md`  
+  - `engineering_manager_position_statement` → writes `output/manager.md`  
+  - `final_alignment_and_resolution` → writes `output/conclusion.md`
+
+- **Crew orchestration**: `crew.py`  
+  - Defines the `LeadershipCalibration` crew and wires agents + tasks.
+
+- **CLI entrypoints**: `main.py`  
+  - Provides `run`, `train`, `replay`, `test`, and `run_with_trigger` functions.
+
+- **Gradio app**: `app.py`  
+  - Defines a small UI where you can type a topic and watch the debate stream.
+
+Generated outputs are written under an `output/` folder (either in the package or project root,
+depending on how you invoke the crew).
+
+---
+
+## Running the Gradio UI
+
+The interactive web UI is defined in `src/leadership_calibration/app.py` using Gradio.
+To launch it:
+
+```bash
+uv run app.py
+```
+
+Then:
+
+1. Open the local URL printed by Gradio in your browser.
+2. Enter a **debate topic** (e.g. “Should we migrate to microservices?”).
+3. Click **Start Debate** to begin streaming the discussion.
+4. Use **Cancel Debate** to stop an in-progress run.
+
+The final markdown output still lands in the `output/` folder and is also shown in the UI.
+
+---
+
+## Customizing the Leadership Calibration
+
+- **Agents** – edit `src/leadership_calibration/config/agents.yaml`  
+  - Update roles, goals, and instructional style (`instructions` block).
+
+- **Tasks & debate flow** – edit `src/leadership_calibration/config/tasks.yaml`  
+  - Change task descriptions, expected outputs, or ordering.
+
+- **Crew wiring & tools** – edit `src/leadership_calibration/crew.py`  
+  - Add or modify tools, change which agent runs which task, or tweak parameters.
+
+- **CLI behavior** – edit `src/leadership_calibration/main.py`  
+  - Adjust default topics, input schema, or logging.
+
+- **UI behavior** – edit `src/leadership_calibration/app.py` and `debate_runner.py`  
+  - Change layout, streaming behavior, or how the debate is rendered.
+
+---
+
+## Support
+
+For more about crewAI:
+
+- **Docs**: `https://docs.crewai.com`  
+- **GitHub**: `https://github.com/joaomdmoura/crewai`  
+- **Discord**: `https://discord.com/invite/X4JWnZnxPb`
+
+This project is a specialized example on top of crewAI for calibrating leadership decisions via structured debate.
+

3_crew/community_contributions/leadership_calibration/knowledge/user_preference.txt

@@ -0,0 +1,4 @@
+User name is John Doe.
+User is an AI Engineer.
+User is interested in AI Agents.
+User is based in San Francisco, California.

3_crew/community_contributions/leadership_calibration/pyproject.toml

@@ -0,0 +1,25 @@
+[project]
+name = "leadership_calibration"
+version = "0.1.0"
+description = "Leadership Calibration using crewAI"
+authors = [{ name = "Your Name", email = "you@example.com" }]
+requires-python = ">=3.10,<3.14"
+dependencies = [
+    "crewai[tools]==1.8.1",
+    "gradio>=6.8.0",
+]
+
+[project.scripts]
+leadership_calibration = "leadership_calibration.main:run"
+run_crew = "leadership_calibration.main:run"
+train = "leadership_calibration.main:train"
+replay = "leadership_calibration.main:replay"
+test = "leadership_calibration.main:test"
+run_with_trigger = "leadership_calibration.main:run_with_trigger"
+
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[tool.crewai]
+type = "crew"

3_crew/community_contributions/leadership_calibration/src/leadership_calibration/app.py

@@ -0,0 +1,87 @@
+import gradio as gr
+import time
+from debate_runner import run_debate, run_debate_stream, cancel_debate
+
+
+def debate_interface(topic):
+    if not topic.strip():
+        return "Please enter a debate topic."
+
+    result = run_debate(topic)
+    return result
+
+
+def test_debate(topic):
+    time.sleep(10)
+    result = f"""**Summary of Both Positions**
+                The debate surrounding {topic}"""
+    return result
+
+
+def handle_button(topic, state, button_label):
+    # If currently running → cancel
+    if button_label == "Cancel Debate":
+        state = cancel_debate(state)
+        return (
+            "❌ Cancelling debate...",
+            state,
+            gr.update(value="Start Debate", variant="primary")
+        )
+
+    # Otherwise start debate
+    generator = run_debate_stream(topic, state)
+
+    # First update button to Cancel
+    yield (
+        "⏳ Starting debate...\n",
+        state,
+        gr.update(value="Cancel Debate", variant="stop")
+    )
+
+    # Stream debate
+    for output, updated_state in generator:
+        yield (
+            output,
+            updated_state,
+            gr.update(value="Cancel Debate", variant="stop")
+        )
+
+    # When finished → reset button
+    yield (
+        output,
+        updated_state,
+        gr.update(value="Start Debate", variant="primary")
+    )
+
+
+with gr.Blocks() as app:
+    gr.Markdown("# 🧠 Leadership Calibration Debate Engine")
+    gr.Markdown("""Enter any topic and watch two Leadership personas (Technical
+    Architect and Engineering Manager) debate and come up with final
+    conclusion. \n\n
+    The Technical Architect is responsible for the technical aspects of the
+    project. \n\n
+    The Engineering Manager is responsible for the people aspects of the
+    project.""")
+
+    topic_input = gr.TextArea(
+        label="Debate Topic",
+        placeholder="Example: Should companies adopt a 4-day work week?",
+        lines=5
+    )
+
+    state = gr.State({"cancelled": False})
+
+    debate_button = gr.Button("Start Debate", variant="primary")
+
+    output = gr.Markdown(visible=True)
+
+    debate_button.click(
+        handle_button,
+        inputs=[topic_input, state, debate_button],
+        outputs=[output, state, debate_button]
+    )
+
+if __name__ == "__main__":
+    app.queue()
+    app.launch()

3_crew/community_contributions/leadership_calibration/src/leadership_calibration/config/agents.yaml

@@ -0,0 +1,73 @@
+engineering_manager_agent:
+  role: Senior Software Engineering Manager & Technical Leader
+
+  goal: >
+    Lead engineering initiatives aligned with business goals while ensuring
+    high technical quality, sustainable delivery, and strong team performance.
+
+  backstory: >
+    You are an experienced engineering leader who has managed high-performing
+    teams in scaling technology organizations. You combine strategic thinking,
+    technical depth, and people leadership. You have led roadmap planning,
+    architecture decisions, incident management, hiring, and performance reviews.
+    You prioritize outcomes over output and sustainability over short-term wins.
+
+  verbose: true
+  allow_delegation: false
+
+  instructions: >
+    When responding:
+
+    - Align technical decisions with business priorities.
+    - Explicitly identify risks, trade-offs, and dependencies.
+    - Separate strategic direction from tactical steps.
+    - Promote ownership, accountability, and psychological safety.
+    - Provide constructive and actionable feedback.
+    - Evaluate scalability, maintainability, and technical debt.
+    - Structure complex responses as:
+        1. Executive Summary
+        2. Key Risks
+        3. Recommended Approach
+        4. Trade-offs
+        5. Execution Plan
+        6. Success Metrics
+    - Ask clarifying questions if requirements are unclear.
+    - Be decisive, structured, and professional.
+
+
+senior_architect_agent:
+  role: Senior Software Engineer & Software Architect
+
+  goal: >
+    Design scalable, resilient, and maintainable systems and deliver
+    production-grade technical solutions with clearly articulated trade-offs.
+
+  backstory: >
+    You are a highly experienced senior engineer and architect who has built
+    distributed systems and operated them in production. You specialize in
+    system design, performance optimization, API design, data modeling,
+    reliability engineering, and operational excellence. You care deeply
+    about clean abstractions, failure modes, observability, and long-term maintainability.
+
+  verbose: true
+  allow_delegation: false
+
+  instructions: >
+    When responding:
+
+    - Clearly frame the problem and identify assumptions.
+    - Distinguish functional vs non-functional requirements.
+    - Define architecture components and boundaries.
+    - Describe data flow and scalability model.
+    - Identify bottlenecks and failure modes.
+    - Evaluate trade-offs (complexity vs scalability vs cost).
+    - Consider security, observability, deployment, and rollback.
+    - When producing code, write production-ready solutions.
+    - Structure complex responses as:
+        1. Problem Framing
+        2. Proposed Design
+        3. Alternatives
+        4. Risks
+        5. Recommendation
+    - Push back on unsafe or fragile designs.
+    - Prefer simplicity unless complexity is justified.

3_crew/community_contributions/leadership_calibration/src/leadership_calibration/config/tasks.yaml

@@ -0,0 +1,62 @@
+architect_position_statement:
+  description: >
+  description: >
+    Topic for debate:
+    "{topic}"
+
+    Provide a structured position from your assigned perspective.
+    Define assumptions, key arguments, risks, and initial recommendation.
+
+  expected_output: >
+    - Core Assumptions
+    - Key Arguments
+    - Risks & Trade-offs
+    - Initial Conclusion
+
+  output_file: output/architect.md
+  agent: senior_architect_agent
+
+
+engineering_manager_position_statement:
+  description: >
+    Topic for debate:
+    "{topic}"
+
+    Respond to the previous position.
+    Critique assumptions and provide counterarguments.
+    Offer alternative framing and recommendations.
+
+  expected_output: >
+    - Agreement / Disagreement
+    - Counterarguments
+    - Broader Impact Analysis
+    - Revised Recommendation
+
+  output_file: output/manager.md
+  agent: engineering_manager_agent
+
+  context:
+    - architect_position_statement
+
+
+final_alignment_and_resolution:
+  description: >
+    Topic for debate:
+    "{topic}"
+
+    Synthesize both perspectives and produce a balanced final resolution.
+    The resolution should integrate analytical rigor and human-centered reasoning.
+
+  expected_output: >
+    - Summary of Both Positions
+    - Key Tensions Identified
+    - Integrated Balanced Resolution
+    - Contexts Where Each View Dominates
+    - Final Recommendation
+
+  output_file: output/conclusion.md
+  agent: senior_architect_agent
+
+  context:
+    - architect_position_statement
+    - engineering_manager_position_statement