codebase-spec-extractor

To extract engineering specifications from a codebase, run Bash scripts to scan project structure, inventory elements, and generate a spec skeleton. This produces documentation of architecture, data models, dependencies, and data flows, enabling system replication without original source code.

Documenting legacy system architecture is best achieved by generating an implementable engineering specification from the codebase. The skill inventories dependencies and data flows, creating a verification workflow so teams can reproduce behavior from documentation alone.

Yes, you can generate a project specification skeleton using Bash and standard Unix utilities. The skill requires these foundational tools to locate repository structure, with optional ripgrep support for faster scanning of dependencies and data flows.

Knowledge transfer across tech stacks without source code is done by extracting complete engineering specifications. The skill analyzes the original codebase to produce a skeleton and inventory, allowing teams to reproduce behavior in a completely different target stack.

Inventory repository dependencies and data models by running the element inventory script on the project path. This scans the codebase to identify architecture and data flows, outputting a markdown file that defines the implementable specification for cross-stack replication.

Generating engineering specs from code requires Bash and Unix utilities to scan the original repository, meaning you need initial access to the codebase to build the inventory. The resulting specification enables later replication without source, but the extraction phase requires direct file access.