Repair shop
Chaco Shop
- 15modules delivered: work orders, online quotes, inventory, cash register, portal and more
- 267automated tests guard the system on every change
- 100accessibility score and 0 axe errors on the website
Engineering in numbers
How we build software, in figures
Everything comes from our own working tools: the project board (Forja), the quality gate that runs before every push, git, and Lighthouse and axe audits. Measured on October 20, 2026.
How we work, in numbers
Our two organizations run on Forja, our own engineering platform: every piece of work is a task with an owner and a closing note, every change goes through a quality gate, and artificial intelligence works under human supervision.
15,620
commits in 90 days, across 39 repositories
git, all branches · 07/22 to 10/20/2026
71%
of those commits declare AI assistance: every change is signed, tested and reviewed by a person
11,123 co-signed · 10/20/2026
78%
of quality gate runs pass green on the first try; what doesn’t pass doesn’t ship
75 of 96 runs · 10/09 to 10/20/2026
163
tasks closed across this week’s 22 active sprints
Forja · 10/20/2026
179
updates published to every machine on the team: every important change gets announced
Forja · 10/20/2026
1,164
knowledge notes: what we learn gets written down
Both organizations’ vaults · 10/20/2026
500
reusable code pieces catalogued: we don’t rewrite what already works
REUSE catalog · 10/20/2026
587
API routes and 8,165 tests in Forja, the platform we work with and also build
Forja codebase · 10/21/2026
Automated tests per project
Every change goes through these tests before reaching production. These are the ones passing today in each project’s quality gate.
8,065
in total
Each project’s CI · 10/20/2026
- weSmile7,030
- Personajes Mágicos451
- Chaco Shop267
- PeakCore (this website and its CRM)197
- Lucie Bienes Raíces60
- Portfolio60
The quality gate
Before every push it runs lint, types, tests and the full build. If anything fails, the change doesn’t ship.
56
gate runs in one week
CI logs · 10/13 to 10/20/2026
86%
passed green on the first try
48 of 56 runs
2
code errors stopped before reaching a client
The other reds came from the environment, already explained
219 s
average time of a full run
Median 203 s
Quality, measured
95-100
on Lighthouse for this website, mobile and desktop
Production build, median of 5 · 10/21/2026
96 → 100
accessibility of the Chaco Shop website after the audit
Lighthouse 12, median of 3 · 10/17/2026
2 → 0
serious accessibility errors (axe) on the same website
axe-core 4 · 10/17/2026
330 → 247
KB is the Chaco Shop home page after optimizing it
Audit · 10/17/2026
Pace of work
1,169
commits in 90 days
git · 07/22 to 10/20/2026
273
commits in the last 30 days
git · 09/20 to 10/20/2026
6
projects with an active sprint today
Forja · 10/20/2026
Each case, in numbers
Events and entertainment
Personajes Mágicos CR
- 24tasks delivered across build and improvements
- 451automated tests guard bookings, pricing and finances
Real estate
Lucie Bienes Raíces
- 3languages: Spanish, English and French
- 14tasks delivered across build and improvements
- 60automated tests guard the website and its CMS
How we measure
- Forja is our board: every task has an owner, a date and a closing note that says what was verified and where.
- The quality gate runs in Docker on our machines before every push to GitHub, with the same tests as the project.
- Lighthouse and axe run on the production build, using the median of several runs, and the date is recorded.
What we don’t measure yet
The business impact for each client —orders handled faster, bookings without overlaps, hours saved— lives in their own systems. We publish it only with their permission, and starting with the next project we measure the baseline before delivery to show the before and after.
Let’s talk about your project