feat: adding word actions, not done, and lots of failing tests #4

Merged
azpect merged 15 commits from feature/word-motions into master 2026-04-04 11:43:47 -07:00
Showing only changes of commit 7ba94eaeea - Show all commits

25
qodo.md Normal file
View File

@ -0,0 +1,25 @@
### The Core Commands
* `/review`
Asks Gemini to read the entire PR diff and provide a structured summary, score the PR, identify potential bugs, and suggest high-level fixes.
* `/describe`
Automatically rewrites the PR's title and description based on the actual code changes. (Great for when you just want to push code and not write documentation).
* `/improve`
Scans the code and provides actionable, copy-pasteable snippets to improve the code, focusing on performance, security, and best practices.
* `/ask "<your question>"`
Turns the PR comment section into a chat window. It uses the PR diff as context. Example: `/ask "Did I properly handle the null pointer edge cases in the new database function?"`
### Specialized Tools
* `/test`
Asks the AI to generate unit tests specifically tailored for the new or modified code in the PR.
* `/update_changelog`
Automatically drafts an update for your `CHANGELOG.md` file based on the PR's contents.
* `/generate_labels`
Analyzes the code changes and recommends appropriate labels for the PR (e.g., `bug`, `enhancement`, `refactor`).
* `/help`
Forces the bot to reply with a quick cheat sheet of all available commands and usage instructions in case you forget them.
### Pro-Tip: Steering the AI
You can actually pass arguments directly to the commands to give Gemini specific instructions for that specific run.
For example, if you want a review but want it to be hyper-paranoid about security, you can type:
`/review --pr_reviewer.extra_instructions="Focus heavily on potential security vulnerabilities and SQL injection risks."`