RepoLens

Evaluate & compare

Score repos against your own rubric, grade their docs, put any N side-by-side, and export a decision matrix.

The verdict is RepoLens' opinion. Evaluations are yours — a structured way to score, compare, and defend a choice when "it felt right" won't cut it.

Score against your rubric

Press e on a focused card (or click the badge) to open the evaluation panel: 1–5 stars per criterion. The card then carries a badge with the weighted average across the criteria you've scored.

The default rubric is three equally-weighted criteria:

CriterionWhat you're scoring
DocumentationIs it actually explained, or just published?
Type safetyHow much does the type system catch for you?
MaintenanceActive, responsive, and unlikely to rot?

It's your rubric, though — ▣ Edit rubric criteria in the command palette lets you rename, reweight, or replace them (up to six). Once repos are scored you can Sort: Eval score or filter to ▣ Show: Evaluated repos only.

Grade the docs, A–F

Docs Quality gives a repo's documentation a letter grade across six weighted dimensions:

DimensionWeight
README completeness25%
Quickstart / getting started20%
Code examples20%
API reference15%
Changelog10%
Contributing guide10%

The weighted total maps to A (90+) through F (under 40), with a plain yes / partially / no on whether the docs are good enough to onboard against.

Compare repos side-by-side

Two ways in:

  • Quick 2-up — hit on two cards to drop them into a head-to-head.

  • N-way matrix — turn on Select, tick 2–10 repos, and press ⊞ Compare. Each repo becomes a column; the rows line up everything that matters:

    Fit · Fit delta · Health · Stars · Language · Decision · Eval score · (each rubric criterion) · Capabilities · Note

Want a tiebreaker? ✦ Ask AI runs a head-to-head and returns a winner, the reason, and the trade-offs. Export the whole table to CSV or Markdown straight from the modal.

Export a decision matrix

When the call has to leave RepoLens — into a doc, a spreadsheet, Notion, or a PR — the command palette exports your full visible library as a structured matrix:

  • ⊞ Export Decision Matrix (CSV) / (Markdown) — every column: fit, health, stars, language, decision + date, eval score, each rubric criterion, the eval note, and your personal note.
  • Lighter options sit beside it: Export visible repos, Export Library (grouped by fit), and Export Digest (JSON / CSV) for a minimal machine-readable dump.

Everything you've decided, scored, and noted comes along — so the artifact you hand off carries the reasoning, not just the ranking.