Apply output rules that turn coding assistant responses into immediate execution steps
Anyone has likely felt frustrated when a coding helper starts with 'Good question' and delays the core answer. This repository applies an output style that removes such awkward introductions and unnecessary explanations, showing you exactly what to do next. Even without an ADHD diagnosis, these rules help you grasp answers quickly when you are struggling to focus.
Ten rules that change the structure of responses
This tool provides ten clear rules for coding assistants to follow when writing answers. The first rule is to always place the next execution step at the very beginning. Tasks requiring multiple steps are listed with numbers, and only one specific next step is presented at the end. Unnecessary asides and off-topic content are suppressed, and the current state is reiterated in every turn. Time estimates are given as specific numbers in minutes rather than vague terms like 'a little.' Additionally, lists are limited to five items, and introductions, summaries, and closing greetings are completely omitted.
Installation and application
Installation is very simple. You only need to copy and paste the repository's installation instructions into the coding assistant's command input field. These instructions reference the content included in the repository's AGENTS.md file, allowing the tool to apply the necessary settings on its own. Once installation is complete, restart the coding assistant and invoke the skill again to apply the new output style. If you have an existing version installed and want to use your forked version, you must first remove the existing plugin and then add your repository to the marketplace.
Changes in output style
When these rules are applied, the coding assistant's answers change significantly. Previously, an answer might start with 'There are several elements to the authentication flow' and list a long background explanation, but now it starts directly with a command to run npm install jsonwebtoken@latest. It specifies file paths and line numbers, presents the code snippet to modify, and shows the test execution command all at once. Readers no longer have to search for the core point in long sentences; they can simply follow the presented steps. This improves efficiency, especially when fixing specific bugs or adding features in complex codebases.
Things to check before applying
This repository is released under the MIT license, so it can be used freely. However, these rules adjust how the LLM writes its answers; they do not change actual code logic or system performance. Therefore, developers must still review the accuracy and security of the code itself. Additionally, because these rules are optimized for specific coding assistant environments, it is difficult to expect the same effect in other tools. It is best to test directly in your own work environment whether this output style helps your productivity or if it feels like it provides insufficient information.
