Tech Skills Framework

Problem

Technical notes, papers, and meeting discussions are often scattered across files: a term query has no clear lookup path, an uploaded PDF is not structurally processed, and cross-topic insights are hard to reuse. This skill organizes a technical knowledge base into L1 through L5 layers and standardizes query, extraction, linking, and archiving workflows.

How It Works

• Term lookup: search L5/kbio first, then fall back through L4 to L1; if nothing is found, say the collection does not contain it.
• Cold-start onboarding: use fetch to read a PDF, then produce L2/kbmd knowledge notes, L3/kbg entity-relation records, L4/kbw term cards, and L5/kbio QA pairs.
• Connection and output: use L3/kbg to find cross-domain gaps and suggest research directions; optionally turn a discussion into a two-host podcast script named YYYYMMDD_podcast_[topic].md and archive it to L5/kbio.
• Framework reuse: copy this SKILL.md, replace the knowledge base name and domain hints, and instantiate a new domain-specific knowledge skill.

Boundaries

It expects an ima shared knowledge base with L1-L5 folders plus search, fetch, file_write, and upload_file.py. If the folders do not exist, they must be created or managed through a knowledge-base setup flow. It avoids search(source="web") by default unless an in-knowledge-base file explicitly references an external link and the user asks to access it.