Repository Dashboard

The Dashboard is the first tab of every repository and the only one that scans nothing itself. It reads what every other scan has already found and answers one question: what should I look at first?

The tab

Codebase overview

The panel at the top counts the problems worth naming — dead code segments, missing test coverage, and files that have never been analysed — then points at one file to start with. The Fix first row always names a file that has real findings on it, never merely the largest one.

The four metrics

• Dead Code — the visible flag count, matching the Dead Code tab exactly.
• Health Score — the composite 0–100 figure. See the Health Score page for the formula.
• Tech Debt — an hours range estimated from the open findings.
• Test Coverage — the share of source files that have a matching test file.

Each card carries a sentence derived from its own band, so the number never sits alone: 44 flags reads “High — immediate attention”, and 6% coverage reads “Low — add coverage”.

Top risk files

Files ranked by a score combining their finding count and their size, with the worst row tinted. Each row links straight into the Dead Code tab for that file. Paths are shown in full rather than as basenames, because a repository with twenty-seven route.ts files makes basenames useless.

The lower panels

Recent activity and coverage gaps

Recent activity lists the last analysis runs with their flag counts and dates. Coverage gaps names the files that have never been analysed at all — the count is the true total, not the length of the sample shown under it.

Detailed analysis

Largest files and a language breakdown by extension. Both are read from the index, so they are available as soon as the repository finishes indexing.

Data & AI

Two repository-wide maps live here rather than on the Code Map, because they answer “what should I look at across this codebase” rather than “what does this file do”:

• Data Flow — where request input enters your routes and where it lands, and which routes check it.
• AI Surface — which models your code calls, from where, and whether the cost of each call is tracked.

Team

A compact Key Person Risk panel reading the same ownership scan as the Key Person Risk tab. Scanning in either place fills both.