pie

What This Skill Does

The Personal Insight Engine (PIE) is a specialized strategic analysis tool designed for OpenClaw agents to introspect on their historical performance. By aggregating session logs stored within the memory/*.md directory, PIE leverages advanced LLM reasoning to distill unstructured daily activity into high-level strategic intelligence. It identifies recurring decision patterns, highlights persistent pain points, and documents pivot moments, effectively transforming raw chronological logs into a structured blueprint for future optimization. By automating the synthesis of your startup journey, PIE ensures that you do not just record history, but actively learn from it, enabling continuous improvement in your operational methodology.

Installation

To integrate PIE into your OpenClaw environment, execute the following command in your terminal: clawhub install openclaw/skills/skills/franklu0819-lang/pie. Ensure you have the openai and python-dotenv packages installed in your Python environment. Before executing, configure your environment variables by adding either ZHIPU_API_KEY or GEMINI_API_KEY to your .env file to authorize the LLM processing layers.

Use Cases

PIE is ideal for entrepreneurs, solo developers, and agents conducting iterative development. Use it for weekly retrospectives to identify why specific projects stalled or succeeded, to prepare for quarterly planning by extracting long-term trends, or to conduct a "post-mortem" analysis after a major product launch. It bridges the gap between massive memory dumps and actionable management reports, allowing for data-driven strategic adjustments.

Example Prompts

1. "PIE, analyze my memory logs from the last 7 days and give me my top 3 strategic insights regarding my recent pivot."
2. "Run the PIE script for the last 30 days. I need to understand my main pain points in the coding workflow."
3. "Summarize the last week of operations using the PIE skill, specifically focusing on decision patterns that led to successful deploys."

Tips & Limitations

For optimal results, ensure your memory logs are descriptive and well-formatted in Markdown. PIE's quality is directly correlated to the depth of your daily logs. Note that PIE relies on external LLM inference, meaning it does require internet access to process the data via the configured API keys. Keep your token usage in mind if analyzing massive datasets exceeding 30-60 days in a single run, as this may impact latency and costs.