adlc-levelup-trace

To generate an AI session execution trace, run the trace generator after completing your feature implementation. It produces a structured trace.md file containing a narrative and technical sections summarizing decisions and solutions.

Producing an AI session trace requires tasks_meta.json and an active feature branch. The generator also validates prerequisites including spec.md, plan.md, and tasks.md before creating the trace.md file.

An AI session trace is a structured document capturing implementation metadata and feature artifacts. It contains a three-part narrative and five technical sections documenting execution context, patterns, and commit references.

You need to create a session trace after implementing a feature to capture decisions and learnings for future reuse. It automates the capture of execution metadata, enabling reproducible learning and auditing.

The trace generator is designed to run after completing /implement. It relies on implementation metadata and feature artifacts produced during that phase to validate prerequisites and generate the trace.md file.

The generated execution trace is saved as specs/{BRANCH}/trace.md. The directory structure corresponds to your active feature branch, organizing traces alongside other specification documents.