... hours. Every test had passed. That's the part I keep thinking about. The tests run against development, not production. They check that the code does what it should. Nothing checked the order things shipped in. So a green suite told me the code was fine. It said nothing about the deploy. If you ship with AI agents, what checks the deploy itself? Migrations, order of steps, anything that only exists in production? Comment from Neelagiri Chettiyar(@neelagiri_chettiyar ... ... only thing that reliably caught this for us was a person still pressing deploy, not the agent, and then reading the live page back after instead of trusting the run that said it was fine. the green

... pipeline looked amazing for a minute. Then I actually opened the test files and realized it was mocking out every single dependency and basically just asserting that true. It feels like the models are trained to prioritize getting a green checkmark over actually testing real failure states. Have you found a specific prompt trick or rule in your system instructions that forces the agent to write tests that can actually fail ...

... Nico s thread this week (green tests, 5.5 hours down) reminded me of a smaller version of the same problem on the B2B AI product I manage. The scene: a data preparation job finishes and the row turns green

... suggested fix where needed. It s designed to run safely on a live site, without generating attack traffic or sending login attempts over the network. No staging site is needed. One thing I m particularly careful about: a green screen shouldn t imply that your entire website is secure. These tests cover the WordPress layer, not your hosting or CDN firewall. The report makes that distinction clear. I d love to hear from site owners and agencies: When a security ...

... pull requests went from 104 a day to 4. In our benchmark, with 6 to 8 Claude Code agents on one repo, main went red under Git and Jujutsu and tasks silently went missing. With AVP, main stayed green and nothing was lost, in every run so far. Method and numbers: avp.murmell.com One binary: `avp up` in any Git checkout connects Claude Code and Codex. The Git side is there to make the transition easy for teams on Git today ...

... anon-dev building with AI agents green checkmarks used to be my green

... half finished job later and had to figure out why What's yours and did it change how you set things up Comment from Konstantin Tikhaev(@konstantin_tikhaev): An agent wrote a fix and a test that went green together. Later the fix was reverted to isolate another bug, and the test still passed. The assertion only checked a default state that never changed. Comment from Konstantin Tikhaev(@konstantin_tikhaev): @neelagiri_chettiyar  The CI check has to fail ... ... code cannot pass without altering the baseline assertions, it has to fail. Comment from Neelagiri Chettiyar(@neelagiri_chettiyar): @konstantin_tikhaev  that one's worse than a crash. a failure you would go and look at.. a green

... developer. I build by directing AI agents, so tests are how I know something works. I can't check it by reading the code. For a while the green checkmarks were enough for me. They shouldn't have been. Now every bug fix comes with a test. Before I accept it, the bug gets put back on purpose and the test has to fail. When we started doing that, we found tests that had been passing on broken code ... ... whole time. Someone on Indie Hackers put it better than I could: > a green

... here s what happened. I spent over a decade in IT project management, and I kept seeing the same thing: a project reports green on schedule and budget, leadership feels good, and months later it misses the outcome it was approved for. I wanted to know if it was just me. So I interviewed 20 senior leaders in 8 countries. The research was published in PM World Journal. 15 of the 20 described the same pattern without being prompted ... ... called DaytaBrain. I m looking for 50 people to try it free for 30 days on 2 real projects, every feature included, and tell me honestly what s missing. Would love to hear if you ve seen a green

... loosening it instead of fixing the code. One example: it replaced a strict deep equality check with expect.anything() on an error payload. The test still exists, still runs, still passes. It just stopped asserting anything. The build stayed green for days until users hit silent null-pointer crashes. The same can happen with CI workflows and migrations. If the agent can edit them, it can make the gate disappear. One cheap guardrail I ve seen: a CODEOWNERS rule requiring human ... ... review on tests/, migrations/ and .github/workflows/, so any diff there needs approval no matter how green

... most expensive habit in the whole loop. The code and the test come out of the same reading of what you asked for. If that reading is wrong, both are wrong in the same direction and the suite goes green, which is worse than having no suite at all, because now you've got a receipt. We hit a version of this on Arteza with a balance that went stale. The check was green

... Netfox's Wi-Fi tool puts a padlock beside every network: green and closed when the traffic is encrypted, red and open when it isn't. Version 0.29.0 added a third case, Enhanced Open (OWE): a network you join without a password whose traffic is still encrypted. It deserves the green

... works for people here, with or without tools. Comment from Konstantin Tikhaev(@konstantin_tikhaev): @syef_sid_ali_benkrief  Workflow files and test suites have to be locked with CODEOWNERS. If the agent can rewrite its own evaluator, a green build is completely meaningless. Comment from Prajwal(@prajwal_paloju1): One check I d add is a fresh-browser run: can someone with no saved session find the main action, complete it, and get an honest success or error? A green

... build WordPress sites and something started bugging me last year. Pages that ranked fine on Google, all green in Yoast, were nowhere when I asked ChatGPT or Perplexity the question the page was written to answer. It cited someone else every time. Went looking for why and found a 2024 Princeton-led study that measured what actually makes an AI engine cite a source. It isn't keyword density or any of the things the green

... idle event no longer overwrites a newer running state. OpenCode 2 and Grok Build also received updated integration paths for execution, input requests, failures, and completion. v0.10.9 puts compact status icons in workspace order: gray for idle, animated green for running or thinking, red for errors or approval requests, and a green

... mention each other, one of which quietly depends on the other, one renaming a field the other one reads, one changing a file format the other one parses. Both agents ran the tests before finishing. All twenty worktrees were green. Git stopped five of the ten merges with conflict markers. It passed the other five as clean, and the merged tree failed the combined suite in all five. The failing tests were always the second agent's own, unchanged, green

... suggestions. The differentiation always turns out to live one layer down, in which single signal the product picked to be good at and that's also where the category quietly fails. Look at what's actually out there. One competitor nails color harmony genuinely good color theory, matches pieces that visually work together. Another does a chat-style interview to build a style profile. A third tracks wear history and layers in weather data so you stop repeating the same ... ... interact. A color pairing that's textbook-correct (complementary hues, right off the color wheel) can still clash because it ignores undertone: burgundy is a cool red, rust is a warm red, and pairing either with the wrong-temperature green

... reads its page list out of the application rather than from a hand picked array, which was already a fix once, because the hand picked version had left three legal pages unchecked. Seventeen pages. Green for weeks. Here is what a stranger can reach that is not among the seventeen: sign in, sign up, forgot password, reset password, delete account, the first screen of the signup wizard, and the page a customer of my customer lands on after paying ...

... chose not to answer. There was nothing to see That's fixed in 0.22.1. The more useful half is the test. Every check in that suite compared against numbers typed into the suite itself, so the whole thing stayed green on a machine the table had never heard of which is precisely the broken state. One of them even asserted that the next macOS was unknown, so it would have passed by agreeing with the bug. There ... ... check that asks the Mac it's running on whether Netfox can name its system. Green

... week I shipped a release and checked it twice. The registry had the new version. The pinned CDN link served the new file. The link people actually install from was still serving the old one. No error anywhere. A green check only tells you about the thing it checked. What's the one thing you still check yourself, even when the agent and the tests both say it's fine ...

... That number is well guarded. There is a unit test named after it that asserts six, and a second one checking that six plus the two background workers fit the host's memory at the per-worker ceiling. Both green, and green

... gates in my own pipeline. Introduce exactly the defect each one claims to catch, confirm that specific check goes red, revert. First run: four caught, three apparently broken. Second run, after I aimed the mutations properly: all three still green. I was about a paragraph into writing up which of my own checks were dead. Then I looked at the script. I was filtering its output with grep -iE 'FAIL' piped into grep -vE '^\s+'. Every line my gate prints ... ... exercise about finding things that cannot report failures. Third attempt, with one change at the top: before trusting any result, run the detector against a state I already know is broken, and abort if it comes back green

... every team I've been on, status colors get gamed before they get reported. "Green" often doesn't mean on track. It means "I don't want the conversation that comes with yellow." Nobody wants to be the one yellow in a room full of green

... asked Faruk what his CLI's check command exits when it cannot read the sensor at all. If that is 0, then "within limits" and "I could not look" are the same exit code, and a cron job sits green while the machine cooks. He had already handled that one. What came out of it afterwards was better: a watch could evaluate a cached reading after the provider underneath it had already failed, and report healthy from it. That ... ... silence. That is a green

... files (which we have already installed while on-boarding) and learn how to connect to SKI. It will run a python script and connect to SKI. Once this happens, and the connection is established, the project name and a green dot will come up on the SKI widget or notch. The green