FAQ: Five Myths About Agent-Installed Dependencies
Did the agent actually install that package? The chat says yes. The next import says no. I keep seeing this exact gap. The log looks green. The environment is not. This FAQ is not a model-IQ rant. It is a state problem. Files versus sentences. Why this keeps biting people Agent loops narrate work in English. Package managers speak exit codes and hashes. We mix those two languages constantly. Then…
Fact 1: The article addresses common misconceptions about dependencies installed by agents. It emphasizes the importance of verifying the actual installation process rather than relying solely on chat logs.
Fact 2: The script provided in the article, named verify-install.sh, is intended as a checklist template for verifying installation details before and after an agent's actions. It compares file hashes and import results to detect discrepancies.
Fact 3: Myth 1 suggests that seeing a successful "Done" message from the agent, such as pip install pydantic, does not necessarily mean the package was properly installed. The article instructs readers to check the output of pip show and a direct import to confirm the installation.
Fact 4: Myth 2 highlights the discrepancy between package names on PyPI and the actual import names on a specific host. The article advises collecting metadata for both the installed package and the imported module to distinguish between them, using commands like pip show and importlib.util.find_spec.
Fact 5: Myth 3 points out that a successful pip install may not necessarily update the lockfile, which can lead to inconsistencies between the installed dependencies and the specified versions in the lockfile. The article demonstrates how to compare lockfile hashes before and after an installation to detect any changes.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
