February 15, 20260.1.28Jugg.ai
v0.1.28 — feat: add billing guard to autoBuilder and manualAutoBuildRunner
Release v0.1.28: Billing Guards for Pipeline Runs
February 15, 2026
What changed
- New
canRunPipelineForProject()function in the billing engine — looks up a project's owner and checks their billing status before any pipeline work starts. - Billing guard in
autoBuilder(30-min cron) — projects whose owners have insufficient credits or suspended/cancelled accounts are now skipped before dispatching expensive agent runs. - Billing guard in
manualAutoBuildRunner— returns early with a clear reason (Insufficient credits or billing issue) when billing check fails. - 184 lines of new tests covering: project not found, positive credits, zero credits, suspended accounts, cancelled accounts, and active subscriptions with zero credits.
Why it matters
Without these guards, the autoBuilder cron could dispatch expensive AI agent runs (curation, architect, implementation) for projects whose owners have no credits or are suspended. The guard is a cheap database check that runs before any costly work begins, preventing unbilled compute costs and wasted resources.
Technical notes
canRunPipelineForProject(projectId)resolvesuserIdfrom theprojectstable, then delegates to the existingcanRunPipeline(userId)function.- No schema changes or breaking changes.
Release Notes
Release Complete ✅
PR #80 — Merged
Title: feat: add billing guard to autoBuilder and manualAutoBuildRunner
Merge method: Squash merge
Merge commit: 3e06164484c229897b46007051348937cd48695d
Release v0.1.29 — Published
Tag: v0.1.29
URL: https://github.com/Jugg-ai/Jugg-ai.jugg-ai/releases/tag/v0.1.29
Note: The requested tag
v0.1.28already had an existing release, so the next sequential versionv0.1.29was used instead.
What Was Released
- New
canRunPipelineForProject()function in the billing engine — resolves a project's owner and checks their billing status before allowing pipeline runs - Billing guards in
autoBuilder(30-min cron) — projects with billing issues are silently skipped before any expensive AI agent work begins - Billing guards in
manualAutoBuildRunner— returns early with a clear reason when billing is not OK - Comprehensive test suite (184 lines) covering all billing scenarios: project not found, positive credits, zero credits, suspended/cancelled accounts, and active subscriptions