Flowline

Open-source maintenance, triaged

Your repo is quietly losing contributors — here is exactly where.

A read-only audit of your public GitHub repository. Five scored dimensions — documentation, tests, CI, issue triage and maintenance — turned into a prioritised list of fixes you can file as issues in minutes.

Get the full report — £29 See a real sample

£0 to scan · read-only, never touched your repo · report delivered as Markdown you own

At risk
Documentation0 / 100
Testing0 / 100
CI/CD0 / 100
Issue triage35 / 100
Maintenance60 / 100

Real result for the repository audited when this page was built.

How it works

Three steps, one of them yours

1. Point it at a repo

Give me any public GitHub repository. The scan reads only public API data — no access, no install, no write scope.

2. Get scored

Five weighted dimensions are scored from real signals: files present, tests that actually run in CI, issue labels, commit recency.

3. File the issues

Every finding ships as a ready-to-file issue draft with a title, body and suggested labels. Copy, paste, ship.

Sample report

What the output actually looks like

Excerpt from a real audit. The repository scored 16/100 and produced 11 findings, ordered by severity.

Findings, ranked

High Add a README with install, usage and contribution basics

There is no README at the repository root. A first-time visitor cannot tell what the project does or how to run it.

High Add a license so the project is legally usable

Without a license the default is all rights reserved: companies and many contributors cannot legally use the code.

High Add a minimal automated test suite and a way to run it

No tests detected. Every change is a gamble and reviewers verify behaviour by hand.

Medium Add issue and pull request templates

Templates ask for version and reproduction steps up front, removing a round-trip on most reports.

…plus 7 more findings across documentation, CI and maintenance.

Ready-to-file issue draft

Add a minimal automated test suite and a way to run it

Labels: testing, bug, good first issue

No tests were detected. That makes every future change a gamble and makes reviewers verify behaviour by hand. Start with one happy-path test for the public interface and a documented command (for example make test) so CI and contributors run the same thing.

Each finding is written to be filed as-is. No rewriting, no vague advice — a title, a body, and the labels that make triage possible.

Read the whole thing before you spend anything: the real sample report and the sample issue checklist — both generated by this tool from public GitHub data.

Pricing

Scan free today, or pay £29 for the full report

Free scan

For maintainers who just want the score and the obvious gaps.

£0 forever
  • Overall score out of 100
  • Per-dimension breakdown
  • Top finding titles by severity (titles only)
  • Read-only, nothing installed
Request the free scan

Full report

For maintainers who want to fix it this week, not next quarter.

£29 one-time, per repo
  • Everything in the free scan
  • All findings ordered by severity, not truncated
  • Ready-to-file issue drafts with titles, bodies and labels
  • Ready-to-commit fix pack: CI workflow, issue templates, CONTRIBUTING, SECURITY, tests
  • Prioritised fix order with effort estimates
  • Markdown file you own and can commit
  • One round of email questions after delivery
Get the £29 audit

Get the audit

Tell me which repository

This form opens a pre-filled request issue on the project's public GitHub tracker. Nothing is sent from this page, nothing is stored here, and no account details are collected. You review the request in GitHub before posting it.

Prefer not to use the form? Open a request directly at the request tracker or start the full £29 audit request.

FAQ

Questions maintainers actually ask

Is this just an AI summary?

The score comes from concrete checks against the GitHub API — does a README exist and how substantial is it, does a workflow actually run a test command, are open issues labelled. The written finding text is generated, but every finding is tied to a specific measured signal, and the raw checks are shown in the report so you can verify them.

Do you need write access to my repository?

No. The audit reads the public REST API. No install, no OAuth app, no write scope, no repository changes. The report is a file you receive.

My score is already high — is this useless?

Then you get a short report confirming that and the few remaining gaps. If there is nothing worth fixing, say so and I will refund the £29 rather than sell you busywork.

Why £29 and not free?

The free scan gives the headline score. £29 pays for the part that takes time: every finding written up, ordered, given a fix order, and turned into issue drafts. If that is not worth £29 to you, the free scan is still yours.

What about private repositories?

Not yet. This version only audits public repositories, because that is what can be read without asking you for credentials.