skill-idea-miner

To mine session logs for skill ideas, run the Python script `mine_session_logs.py`, which parses Claude Code session logs and outputs scored candidates to a directory you specify.

Scoring skill idea candidates for novelty and feasibility is handled by Python scripts that analyze session logs, assigning scores based on novelty, feasibility, and trading value. Deduplication is also applied to remove similar ideas.

No, you do not need external API keys for LLM calls during session log analysis. The process requires the Claude CLI for LLM calls and standard Python libraries like pyyaml and json for processing logs and scoring ideas.

Yes, you can use Python scripts to automate weekly skill generation by running the mining pipeline, which automatically extracts, scores, and deduplicates new skill ideas from recent coding sessions to build a backlog.

Yes, the session log mining pipeline specifically applies to the development of trading and investing related skills, scoring extracted candidates on trading value alongside novelty and feasibility.