About RepoCritics

RepoCritics is the Michelin Guide for open source. We collect signal from working engineers — the developers actually shipping software with these libraries — and turn it into a score you can trust at a glance.

Why

GitHub stars are vanity. Download counts measure marketing, not quality. A README promises everything; the issue tracker reveals everything. We exist for the gap in between: an honest, ongoing review of OSS as it lives in real codebases.

Two scores

• Critics score — weighted average from invited senior reviewers who have shipped non-trivial production work with the repo. Heavy weight on recency, longitudinal use, and verified review depth.
• Audience score — open community ratings from registered developers, normalized for vote brigading and account-age signals.

How the score is built

The composite score is a blend of:

• Critic ratings (per dimension: API design, docs, perf, ecosystem, stability)
• Audience ratings (capped contribution to limit popularity bias)
• Freshness penalty (stale critique decays)
• Security signal (CVE count, response time, deprecation flags)
• Maintenance signal (issue/PR latency, release cadence)

We publish the formula. The dimension weights are versioned, dated, and reviewable — like any audit-grade rubric.

Who reviews

Anyone can join as an audience member. Critic status is earned: track record of OSS contribution, two reviews accepted by the editorial board, and a real-name public profile. We are slow and skeptical here on purpose.

What we are not

We are not a vulnerability scanner, a download counter, a hype cycle, or a paid promotion vehicle. We do not accept sponsorship from reviewed repos. Anyone may contest a review; the board adjudicates publicly.

Status

🚧 Phase 1 — Skeleton. The site you are reading is the scaffold. Reviews, scoring, and accounts ship in subsequent phases.