Skip to content
Testament

ATPR

Measures how far a test suite can be trusted, with evidence; first target is Actual Budget.

AI Trust and Product Readiness: a pipeline that answers how much leadership should trust a codebase's test suite, and with what evidence. Rather than counting tests or pass rates, it measures what the suite would catch if the code were broken, whether each test constrains behaviour, and whether its past confidence matched the bugs that escaped, scoped to the release diff. It audits its target from outside, never changing it; the first target is Actual Budget's loot-core, pinned to a release tag.

Reporting healthyLast report
Details hidden: private repository

This repository is private, so failure messages, stack traces and source links are hidden. Test names, counts and trends are real.

PassedLatest run · · main · d39492f
175tests · 175 passed · 0 failed · 0 skipped · 0 s
Pass rate, 30 days
100%
Green streak
20runs
Longest20 runs
Time to green
None
No recoveries in 90 days
View this run →

What it’s built with

Built with

Language
  • Node 24
  • TypeScript 7
Libraries
  • Zod 4
  • yaml
Delivery
  • GitHub Actions

Tested with

Runners
  • Vitest 5

How it’s tested

175tests executed in the latest run
  • Integration
    32%
  • Unit
    17298%

Coverage and reports

Reports per run

Job / module / platformTests
test/atpr/node175

History

Pass rate

Default branch · last 30 runs · CI and imported history

All 20 runs passed.

Tests per run

Default branch · last 30 CI runs

Grew from 64 to 175 over the last 20 runs.

Run duration

Default branch · last 30 CI runs (imported history has no durations)

Between 0 s and 1 s over the last 20 runs. Median 0 s.

Last 40 runs

Default branch · run status

All 20 passed.
Latest
  • Passed

Runs and flaky tests