Repository Technical-Debt Analysis

Produces a prioritized inventory of the technical debt of a repository, together with an estimate of what each item costs and what removing it costs.

Two principles drive everything below:

1. Debt only matters where the code is touched. Ugly code that nobody reads and nobody changes pays no interest. A moderate smell in the file that changes every week is far more expensive than a severe smell in a file untouched for five years. Prioritization must therefore be based on evidence from the repository's history, not on aesthetic judgement of a snapshot.
2. A finding without evidence is an opinion. Every finding names a concrete location, shows what is actually there, and states why it costs something. "Could be cleaner" is not a finding.

This skill is read-only: it produces a report, it does not modify, fix, refactor or reformat anything (see "Non-goals").

Completion criterion