auto

name	auto
description	Autonomous task execution cycle: Plan -> Research -> Prototype -> Evaluate -> Report

Auto Execution Skill

Core Philosophy

"Autonomous, domain-expert execution from start to finish."

This skill empowers the agent to take a high-level topic or request and self-manage the entire lifecycle of delivery, from initial planning to final reporting, while adopting a specific expert persona.

Process

1. Role Definition & Prompt Generation

Goal: Establish the expert persona best suited for the task.

1. Analyze Request: Identify the core domain (e.g., "Bioinformatics", "Rust Systems Programming", "Market Analysis").
2. Define Persona: Choose a title (e.g., "Senior Bioinformatics Researcher", "Principal Systems Engineer").
3. Generate Prompt: Create a specific system prompt for yourself for this task.
   • Instruction: "I am now acting as [Role]. My specific expertise includes [Key Skills]. I will approach [Topic] by [Methodology]."
   • Output: Explicitly state this role at the beginning of the execution.

2. Operational Outline

Goal: Create a roadmap for the specific task.

Before executing, generate a step-by-step Execution Outline following this strict structure:

1. Concepts (Các khái niệm): Define key terminology, background research, and core concepts relevant to the topic.
2. Architecture & Flow (Luồng hoạt động/Kiến trúc): Design the system architecture, data flow, or logical process steps (if applicable).
3. Demo & Testing (Demo và kiểm thử): Build the prototype/proof-of-concept and verify it with tests.
4. Evaluation Report (Báo cáo đánh giá): Synthesize findings, test results, and expert conclusion into a final report.

3. Execution Loop

Execute the outline steps sequentially.

1. Concepts

• Goal: Understand and define the problem domain.
• Use academic-search or search_web to define key terms ("Các khái niệm").
• Save notes to notes/ or include in the plan.

2. Architecture & Flow

• Goal: Design the solution structure ("Luồng hoạt động/Kiến trúc").
• Create a plan file results/[YYYY-MM-DD]-[Topic-Slug]/plan.md.
• visual/textual description of how the system/process works.

3. Demo & Testing

• Build:
  • Create folder results/[YYYY-MM-DD]-[Topic-Slug]/demo/.
  • Use prototyping skill to build there.
  • Write code (app.py, requirements.txt).
• Test:
  • Run: Execute the prototype ensuring no errors.
  • Verify: Check if it demonstrates the concepts and flow defined earlier.
  • Fix: Iterate until the demo is stable.

4. Evaluation Report

Goal: Deliver a comprehensive report ("Báo cáo đánh giá").

• Create a file: results/[YYYY-MM-DD]-[Topic-Slug]/report.md.
• Structure:
  1. Executive Summary: The "Bottom Line Up Front".
  2. Concepts: Summary of the key concepts researched.
  3. Architecture: Description of the flow/design.
  4. Demo & Results:
     • Links to code/demos.
     • Screenshot descriptions/Logs.
     • Evaluation: Did it work? What are the limitations?
  5. Conclusion: Expert verdict.

Usage Tips

• Be Autonomous: Do not ask the user for permission between every step unless critical.
• Self-Correct: If the prototype fails, fix it. Don't report failure unless unfixable.
• Expert Tone: Maintain the persona throughout the report.