-
Ten thousand venture scripts, zero syntax errors
A node --check pass over every standalone .js and .mjs file under ventures/ found no syntax errors. What the sweep covered and what it cannot see.
-
The slowest CI step was also the one that was failing
A performance reading of the site builder's tests found 71 job-minutes of a 510-minute window, and three failures from one new page that made every run wasted.
-
Why a stair landing could need 192 billion posts
Probing the wall-framing takeoff's stair helpers with hostile numbers found a post counter that would happily order billions of posts. Now it refuses.
-
Why the bid package licence check got a fuzz test
The branded bid package unlocks on a licence key check. A seeded property test now throws junk keys, URLs and server answers at it.
-
How do you test a random schedule forecast?
The wall-framing takeoff's schedule risk tool rolls dice over a builder's duration ranges. Here is how a seeded property test checks it without a real project.
-
Testing the free custom development queue from form to page
End-to-end tests for the free custom development queue found two bugs: form requests filed under the form's header, and preview feedback queued as new requests.
-
Testing the typo catcher in a framing calculator
A property test for the check that catches misspelt plan keys in a framing calculator.
-
Two bugs we could plant in the model-key helper without a test noticing
A seeded-fault run on the bring-your-own-key helper caught 6 of 8 mutants; the two escapes, a default-port edge and a private range, are now guarded.
-
Red team: two over-long post descriptions turned the distribution hub suite red
A red-team run found the hub's tests failing on main because two new posts had descriptions over 160 characters. Fixed, plus a test for one escaped mutant.
-
Red team on the narration tool: 3 of 8 seeded bugs caught, then 8 of 8
Seeding 8 bugs into platform/audio/narrate.py, the tests caught 3. Four new stubbed tests cover model download, loading, pool sizing and an empty script.
-
We seeded 24 bugs in three build scripts; 14 were caught
A mutation campaign on the stories page, the releases tool and the SERP snippet audit: 58% caught, three real gaps in the stories page now closed.
-
The nightly browser checks had not passed in seven weeks. Here is what they found when they ran
A newer Playwright and one job per check revived the nightly browser checks. The first full run found a dead calculator page and 23 pages 2px too wide.
-
Fifteen days of red team: what it broke, and what the blue team fixed
294 red-team runs seeded 3,437 bugs and none is in the code today. The blue team closed 864 test gaps and 89 defects, yet fresh files still miss 3 in 10.
-
Seeding bugs in the shop-rates bar: 4 of 8 got through
A red-team campaign on the shared shop-rates bar caught half its seeded bugs; the escapes included one that would overwrite a shared bid link.
-
We broke eight lines of our own browser script and no test noticed
A red-team run flipped eight lines in the script that points at the invalid field on our calculators. None failed a test. Here is what that found.
-
Red team on the posts-per-day script: 8 of 8 seeded bugs caught, 2 real ones found
We seeded eight faults into platform/ci/post_rate.py; the tests caught all eight. Reading it by hand still found two real bugs.
-
Do the MACRS tables still add up to the cost after rounding?
A 600-asset property test of the ai-bookkeeper tax depreciation code: the years sum to cost exactly, under half-year and mid-quarter rules.
-
We broke our own opt-out on purpose, and one change slipped past the tests
A planted bug flipped the default of the usage opt-out and 7 of 8 mutants were caught; we added the test that pins the missing one.
-
Six of eight seeded bugs in the saved-bid backup got through the tests
A red-team run seeded eight faults in platform/pages/backup.mjs. The tests caught two. The escapes sat mostly in browser-run code no test executed.
-
All 1,326 venture pages load in a real browser with no JavaScript errors
A headless Chromium sweep of every venture page found no uncaught exceptions or console errors, and the sweep is now a script anyone can rerun.
-
A favicon sweep left 15 venture test suites red on main
Running all 1,314 venture suites found 15 failing. The cause: inline SVG tab icons added to pages without updating the tests and generators that check them.
-
Our mutation tester said 'caught nothing' because it ran zero tests
A red-team campaign reported 0 of 8 seeded bugs caught. The cause was a runner that found no pytest-style tests. The fix, and the gaps found after it.
-
A test for ignoring gh-pages that could not fail
Seeding bugs into the unlanded-work checker showed that its gh-pages test passed with the filter removed. Why, and what now guards it.
-
Seeding faults into our CI checkers: 28 of 32 caught, 4 provably harmless
A red-team run mutated four platform/ci checkers 32 times. The tests caught 28; the other 4 cannot change what a user sees.
-
Seeding bugs in the estimated-tax engine: 3 of 8 caught
A red-team campaign on ai-bookkeeper's estimatedtax.mjs found five escaping mutants; five new assertions now catch them.
-
Seeding bugs into the tree-count sync script: 1 of 10 caught
A red-team run mutated three CI scripts. Two caught every seeded bug; sync_tree_counts.py caught one in ten, and now catches seven more.
-
59 valid e-invoices were refused because of a licence comment
Running einvoice-preflight over the ZUGFeRD corpus PDFs found it could open 40 of 126 valid files. The cause was a 4 KB sniff window.
-
We broke our own WordPress plugin builder 8 ways and the tests noticed 2
A red-team run seeded faults into five CI and distribution scripts. Catch rate was 76%, lowest on the WordPress embed builder (2 of 8).
-
Three N26 exports imported with blank descriptions. Here is how we found out
Reading open-source bank plugins' column layouts turned up a real gap in our statement reader: N26's Italian, French and Spanish files lost every payee.
-
The official Peppol negative-invoice example was rejected by our e-invoice checker
einvoice-preflight on Peppol BIS 3.0 and ZUGFeRD community sample invoices found five false positives, from negative invoices to a purchase-order placeholder.
-
What happens when an agent types -1 into a calculator
Pages clamp negative inputs but agents call the engines directly. A new probe found 21 calculators that carry a negative into the math; four are fixed.
-
Seeding 8 bugs into the npm publisher: 6 caught, and the 2 that slipped through
A mutation campaign on platform/distribution/npm.mjs caught 6 of 8 seeded faults. Both escapes were in the package.json keyword filter; one test closes them.
-
We seeded 16 bugs into two checks; the tests caught 9
A red-team run mutated the blog front-matter check and the venture-defaults grounder. Seven escaped. Now the guards catch all but one equivalent mutant.
-
Six test suites were red before we touched them
A red-team run found a favicon data: URI tripping link checks in six ventures, and 4 of 8 seeded bugs in the prop-firm calculator slipping past its tests.
-
Red team: the exit code nobody tested
Seeding 16 bugs into two CI tools found one real hole: public_surface.py's exit 1 versus exit 2 was untested.
-
We seeded 8 bugs in the heading-order check; its tests caught 2
A red-team run on platform/ci/check_heading_order.py: six of eight seeded faults escaped, and the new tests catch all six.
-
We tried our BAI2 reader on four public bank files and it miscounted accounts
A bookkeeping parser written from memory of the BAI2 layout met four real-shaped sample files. It read the money right and got one thing wrong.
-
The bank CSV bug that turns 5 March into 3 May
Probing a bookkeeping importer with seven made-up Nordic, Thai and Icelandic bank files: one silently swapped day and month, two mislabelled columns.
-
We read other projects' OFX files with our bookkeeper. It got 13 right and one wrong thing.
ai-bookkeeper's bank readers checked against sample files from ofxparse and ofxtools: every transaction counted, one character-reference bug fixed.
-
A bookkeeping importer that dropped dividends without a word
Running public OFX sample files through our open-source bookkeeper found a brokerage file where cash income vanished silently. The fix and the test.
-
Checking a payroll deposit scheduler against the IRS's own worked examples
The $100,000 next-day deposit rule in an open-source bookkeeper, tested against the two examples printed in IRS Publication 15 (2026).
-
CRM exports are messier than your test fixture: four fixes in a prospecting agent
A sales research agent that passed its tests on hand-built HubSpot data still misread Excel-style and Pipedrive-style files. What broke and what we changed.
-
A bank header that silently deleted income: Paid In (GBP)
Probing a bookkeeping CSV reader with 41 bank layouts from Firefly III's configs found a header that made money-in rows vanish without a warning.
-
We ran our e-invoice checker on the EU's own example invoices. None passed.
Our free e-invoice preflight failed all 41 official EN 16931 examples. The causes were arithmetic false positives, now fixed.
-
The id that ate the table: a silent bug found by sweeping 2,865 pages
Two elements shared one id, so a script wrote a results table into an input box. Nothing threw. How we found it across 2,865 pages and now block it in CI.