~/blog
Blog
Notes on dependency hygiene, health scoring, and shipping a static-only tool.
2026-06-02·engineering
How we compute the health score (and why it's deterministic)
The exact formula behind reposcan's 0–100 health score, plus why we deliberately avoided ML and weighted heuristics.
read post
2026-05-18·analysis
Anatomy of an abandoned repo
Patterns we see across thousands of scans: the warning signs that show up before a project officially dies.
read post
2026-04-29·philosophy
Why reposcan will never ask you to log in
An opinion piece on tools that should remain frictionless, and what we give up by skipping auth.
read post
2026-04-10·engineering
Living within GitHub's 60-req/hr unauthenticated ceiling
How we batch, cache, and budget API requests to make reposcan feel instant without a backend.
read post
