buildOne report1 publisher Chainguard runs coding agents in remote microVMs and proposes signed attestations as the merge gate
Chainguard has launched two beta products, Workspaces and Checks, that run coding agents in remote microVMs and move CI into the dev loop. Keeping agents off laptops that hold corporate credentials costs teams a dependency on Chainguard's execution and signing system, InfoQ reported.
Reality
- Evidence35
- Adoption
- Not scored
- Hype gap
- Not scored
- Incentives
- Not scored
- Confidence45
buildOne report1 publisher A one-commit checkout and a trailing || true let 41 PRs skip tests under green CI
GitHub Actions' default one-commit checkout let 41 pull requests skip their tests over three weeks while CI stayed green, a dev.to author reports. The author's change-detection step read git's fatal error as an empty diff, and the run that exposed it showed a test job that finished in 4 seconds.
Reality
- Evidence62
- Adoption
- Not scored
- Hype gap+6
- Incentives
- Not scored
- Confidence60
buildOne report1 publisher GitHub is rebuilding its Git storage around the write load agents create
GitHub is rebuilding its Git storage infrastructure as monthly pushes grew 4.9x in a year to 3.35 billion. The strain is on writes, since every push must be durable and consistent before the next agent or CI job can build on it.
Reality
- Evidence45
- Adoption50
- Hype gap+25
- Incentives60
- Confidence55
buildOne report1 publisher Coding agents have made CI the bottleneck at Anthropic and Linear
Anthropic's CI job volume grew 25x in six months as coding agents raised pipeline load, The New Stack reports. Faster runners and test selection cut the cost of each run, yet repo tests still mock the service seams where distributed systems break.
Reality
- Evidence45
- Adoption60
- Hype gap+20
- Incentives50
- Confidence45
buildOne report1 publisher Uber's SubmitQueue tests pending changes against each other to keep six monorepos green
Uber's SubmitQueue builds a speculation tree of pending changes to keep mainline green in monorepos that take more than 65,000 changes a month. At that volume, a one-at-a-time queue would need every build done in about 40 seconds.
Reality
- Evidence40
- Adoption30
- Hype gap+20
- Incentives45
- Confidence55
buildOne report1 publisher Directory batching restored CI throughput after per-file AI commits overloaded an industrial C++ repo
CI and reviewer attention set the pace when one developer's AI tool pushed hundreds of commits into an industrial C++ repo over 15 days. The remaining delay sits in how changes are sliced into commits and CI jobs, a part of the cycle GitHub's sub-30-second Copilot review does not shorten.
Reality
- Evidence30
- Adoption
- Not scored
- Hype gap+20
- Incentives
- Not scored
- Confidence35
buildConfirmed2 publishers Lorin Hochstein rates Buildkite's public write-up above GitHub's on detail, and the first thing his own retelling has to do is teach clusters, nodes and headroom, the layer a platform keeps out of sight.
Publishers:daily.dev · surfingcomplexity.blog
Reality
- Evidence55
- Adoption
- Not scored
- Hype gap0
- Incentives20
- Confidence60
buildOne report1 publisher Capping fetch depth took Linear's slowest CI gate from 94 seconds to 20
Linear rebuilt its CI pipeline after its CTO filed an issue about cost, and the numbers it published show verification volume growing faster than the savings on each test. The largest wall-clock gains came from the jobs that decide what to run.
Reality
- Evidence55
- Adoption25
- Hype gap+10
- Incentives55
- Confidence60
buildOne report1 publisher zerocase refuses green CI runs whose reports show nothing executed
zerocase fails a CI run when its JUnit, TAP, LCOV, Cobertura or ESLint report shows zero executed items, even if the command exited 0. It counts only what ran, so fifty skipped tests fail the gate and fifty failed tests are left to the exit code.
Reality
- Evidence45
- Adoption
- Not scored
- Hype gap+5
- Incentives35
- Confidence45
buildOne report1 publisher canfail breaks the guarded file on purpose to see whether the check notices
Mutation testing needs a function to mutate, so the YAML and Terraform that gate production go untested. canfail edits an anchored string in those files, runs your check and reports whether it failed for the reason you declared.
Reality
- Evidence46
- Adoption9
- Hype gap−12
- Incentives28
- Confidence54
buildOne report1 publisher The proposed fixture scorer accepts any non-empty override token as a human approval
A dev.to scorecard for agent-authored patches digests every file under testdata/ and replays recorded property seeds before pytest runs. The published fragments enforce less than the prose around them describes.
Reality
- Evidence55
- Adoption
- Not scored
- Hype gap+30
- Incentives20
- Confidence70
buildOne report1 publisher Salesforce engineers call flat review time on the biggest pull requests a sign of disengagement
Linear's CI rework and a Salesforce code review postmortem describe different failures in the layer that judges a change. One was fixed with faster runners and a compiler swap, the other by rebuilding how a diff is presented to a human.
Reality
- Evidence52
- Adoption44
- Hype gap+16
- Incentives58
- Confidence50
buildOne report1 publisher Anthropic's CI jobs grew two and a half times as fast as its test suite
The numbers sitting under the 80-percent-of-code headline are the ones a team can budget against: a test suite up 10x and CI jobs up 25x in six months, with a deterministic selection service deciding which tests any change actually runs.
Reality
- Evidence38
- Adoption30
- Hype gap+22
- Incentives60
- Confidence42
buildOne report1 publisher In one worked example, a coding agent waits 93 percent of its cycle on CI
A dev.to design essay argues CI was sized for human review speed and proposes tiering checks by cost and failure probability. Its figures are examples, so the case rests on the failure modes it names.
Reality
- Evidence22
- Adoption
- Not scored
- Hype gap+30
- Incentives25
- Confidence45
buildOne report1 publisher A suite failing one run in four needs each test to flake only 0.3 percent of the time
A dev.to post names five habits that rot a Playwright suite, starting with waitForTimeout used to patch a race condition. Working back from its one-bad-run-in-four figure shows how little per-test failure that takes.
Reality
- Evidence30
- Adoption
- Not scored
- Hype gap+10
- Incentives20
- Confidence55
buildOne report1 publisher Recording the score contract from the parent commit keeps the agent out of its own scorer
A dev.to proposal freezes fixture bytes, runner config and two test floors at the parent SHA, then denies the coding agent write access to the file holding them. Its content hashes hold up better than the path denylist it prints.
Reality
- Evidence45
- Adoption
- Not scored
- Hype gap+15
- Incentives20
- Confidence55
buildOne report1 publisher Split the suite so flaky tests cannot gate an agent's patch
A dev.to write-up sorts every test path into four classes and hashes the scoring ones on main, so a patch that retouches a fixture fails with a different exit code than a patch that breaks a property.
Reality
- Evidence32
- Adoption
- Not scored
- Hype gap+18
- Incentives18
- Confidence45
buildOne report1 publisher A link gate printed PASS over zero files because git ls-files ran against an enclosing repository
A faithful git archive export does not include .git, so the file list came back empty, the command exited 0, and the gate printed the same verdict a clean tree prints. The repair was a third verdict, NOT RUN, that names the root and the command.
Reality
- Evidence55
- Adoption12
- Hype gap−5
- Incentives15
- Confidence60
buildOne report1 publisher Month two of an unreviewed AI test pipeline left 30 percent of generated Playwright tests flaky
A dev.to post-mortem swapped 40 percent of an end-to-end regression suite for LLM-written Playwright tests over six months. Its two published numbers, 85 percent passing first run and 30 percent flaky in month two, count different tests.
Reality
- Evidence30
- Adoption20
- Hype gap+20
- Incentives25
- Confidence45
buildOne report1 publisher Blacksmith enforces its org-only rule by leaving the job queued forever
An unaffiliated 10-trial benchmark clocked Blacksmith's runners at 3.17x and 3.77x faster than ubuntu-latest, with no overlap in the ranges. Reaching that baseline first meant moving the repository off a personal account.
Reality
- Evidence62
- Adoption26
- Hype gap+14
- Incentives24
- Confidence55
Earlier coverage
- ArchUnitTS fails an architecture rule that inspected zero files
Build · September 17, 2026 · One report1 publisher