The dictionary.

LGTM stands for “looks good to me.” Reviewers write it to approve a pull request or code change and signal that it is ready to merge.

Squash and merge combines every commit in a pull request into one commit on the target branch, so the main history gets one tidy entry per change.

A merge conflict happens when Git cannot combine two branches on its own because both changed the same lines, and a person has to decide what stays.

“It works on my machine” means code runs for the person who wrote it but fails elsewhere, usually because the environments differ in some way.

“One last commit” is the small final fix before a branch is ready. It is rarely the last one, and good commit habits keep the history readable anyway.

“Cannot reproduce” means a developer followed a bug report but could not make the bug happen, so it cannot be confirmed or fixed yet.

“Changes requested” is a code review result meaning the reviewer wants specific changes made before the pull request can be approved and merged.

A known issue is a bug or limitation the team has confirmed and documented but has not fixed yet, usually listed in release notes with a workaround.

“Do not merge” marks a pull request that should not land yet, because it is unfinished, waiting on something else, or open for discussion only.

A nit is a small, non-blocking review comment. “Approved with nits” means the change can merge, with a few minor suggestions the author may take.

Rebasing replays your branch’s commits on top of the latest target branch. “Please rebase” asks you to update a branch that has fallen behind.

“Awaiting review” means a pull request is open and waiting for a teammate to read and approve it. Nothing merges until someone does.

A preview deployment is a temporary live copy of a site built from a pull request. “Preview is building” means that copy is not ready yet.

Vibe coding is building software by describing what you want to an AI model and accepting the code it writes, often without reading every line.

“Accept all” applies every change an AI coding tool proposes at once. It is fast, and it works best when you have at least skimmed the diff first.

The context window is how much text an AI model can consider at once: your instructions, the conversation, any files it reads, and its own reply.

Tokens are the small chunks of text AI models read and write. They are how usage is measured, priced and limited, including the size of a context window.

“One more prompt” is the loop of asking an AI for just one more fix. Knowing when to stop, revert or start fresh saves hours and tokens.

An AI coding agent is a model that works in a loop with tools: it reads files, edits code and runs commands to finish a task, then checks its own result.

A hallucination is when an AI model states something false with confidence, like a function, package or setting that does not exist.