Your Agent Says Tests Pass. On Your Laptop. Not on CI.
The agent ran the test command, watched the green checkmark, and said done. You opened the PR. Twenty minutes later CI was red. The agent looked at the failing job, shrugged, and said "works on my machine." It was not being smug. It genuinely had no idea what was on your machine that was not on the runner. This is one of the most common and least discussed failure modes in agentic coding. The…
The article highlights a common issue in agentic coding where an agent's verification loop fails because the tests pass on the agent's laptop but not in the continuous integration (CI) environment. This discrepancy arises because the agent inherits the local machine's environment, including API keys, local databases, and specific package versions, while CI runs in a clean, isolated container.
The author emphasizes the importance of running tests in an environment that closely matches the CI setup, rather than relying on the agent's local environment. Recommendations include running the same command CI uses, ensuring a clean working directory, and avoiding changes to the environment that might mask the underlying issue.
The goal is to ensure that passing tests on the agent's machine genuinely reflect the results CI will produce.
Brief written by urgent.news from Dev.to's own syndicated text. Machine-written — may contain errors; check the original before relying on it.
