# Agent Docstrings - Serve Agents a Clear Table of Contents, Right in Your Code **URL:** \ **Category:** Built for Cursor\ **Tags:** java\ **Created:** [June 29, 2025, 10:42pm UTC](https://forum.cursor.com/t/agent-docstrings-serve-agents-a-clear-table-of-contents-right-in-your-code/110826 "2025-06-29T22:42:25Z")\ **Posts on this page:** 6\ **Page:** 1
**Author:** ![Artemonim](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/artemonim/32/56113_2.png) [@Artemonim](https://forum.cursor.com/u/Artemonim)\ **Post date:** [June 29, 2025, 10:42pm UTC](https://forum.cursor.com/t/agent-docstrings-serve-agents-a-clear-table-of-contents-right-in-your-code/110826/1 "2025-06-29T22:42:25Z")
# Agent Docstrings A command-line tool to auto-generate and update file-level docstrings summarizing classes and functions. Useful for maintaining a high-level overview of your files, especially in projects with code generated or modified by AI assistants. `pip install agent-docstrings` > **[GitHub - Artemonim/AgentDocstrings: A command-line tool to auto-generate and update...](https://github.com/Artemonim/AgentDocstrings)** > > A command-line tool to auto-generate and update file-level docstrings summarizing classes and functions. Useful for maintaining a high-level overview of your files, especially in projects with code generated or modified by AI assistants. ## How to I’m adding this tool to my code quality control script, which the Agent uses for self-verification. I’ll add more information to the project’s README if there are any questions. ## Why? When working in Cursor and similar IDEs, Agents often start reading files from the beginning. And regarding Cursor’s behavior during the script’s creation, in normal mode, the model reads 250 lines of code per call, and in MAX mode, 750 lines. However, I have projects with files over 1000 lines of code, which are not very appropriate to divide into smaller files. And anyway, Agent still have to call reading tools for each individual file. At the same time, the Agent can choose from which line to read the file. It can navigate and surf within your repository. The script literally provides the Agent with the table of contents of the current file, so that immediately after the first read, the Agent understands the entire structure and can read the file from a specific line, rather than trying to get to it (while also potentially making mistakes along the way). In addition to the advantage of quick navigation, the initial docstring also serves as a method to reduce context window usage. For example, if a required method in a 900-line file is on line 856, the Agent will only read lines 1-250 and 856-900, instead of sequentially going to the desired forty lines and filling its context with unnecessary code. ## Features - **Multi-language support** : Python, Java, Kotlin, Go, PowerShell, Delphi, C, C++, C#, JavaScript, TypeScript - **Automatic discovery** : Recursively scans directories for source files - **Smart filtering** : Respects `.gitignore` files and custom blacklist/whitelist configurations - **Incremental updates** : Only modifies files when changes are detected - **Type annotations** : Full type hint support for Python 3.8+ - **CLI interface** : Easy-to-use command-line tool ---
**Author:** ![Artemonim](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/artemonim/32/56113_2.png) [@Artemonim](https://forum.cursor.com/u/Artemonim)\ **Post date:** [July 1, 2025, 5:11am UTC](https://forum.cursor.com/t/agent-docstrings-serve-agents-a-clear-table-of-contents-right-in-your-code/110826/2 "2025-07-01T05:11:13Z")
[Version 1.3.0 has been released](https://github.com/Artemonim/AgentDocstrings/releases/tag/v1.3.0) ---
**Author:** ![Artemonim](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/artemonim/32/56113_2.png) [@Artemonim](https://forum.cursor.com/u/Artemonim)\ **Post date:** [July 2, 2025, 8:12am UTC](https://forum.cursor.com/t/agent-docstrings-serve-agents-a-clear-table-of-contents-right-in-your-code/110826/3 "2025-07-02T08:12:47Z")
[Version 1.3.1 has been released](https://github.com/Artemonim/AgentDocstrings/releases/tag/v1.3.1) - **Feature Added:** Support for merging generated tables of contents into single-line Python docstrings. - **Bug Fix:** Resolved duplication issue when reprocessing files with existing docstrings; improved TOC placement logic. - **CI/CD Updates:** CI now runs on `dev` branch; Codecov limited to `master`; added version bump protection for feature branches. ---
**Author:** ![Artemonim](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/artemonim/32/56113_2.png) [@Artemonim](https://forum.cursor.com/u/Artemonim)\ **Post date:** [July 9, 2025, 9:11am UTC](https://forum.cursor.com/t/agent-docstrings-serve-agents-a-clear-table-of-contents-right-in-your-code/110826/4 "2025-07-09T09:11:36Z")
[Version 1.3.5 has been released](https://github.com/Artemonim/AgentDocstrings/releases/) I hope I’ve finally cleaned up all the bugs… 🫠 ---
**Author:** ![Artemonim](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/artemonim/32/56113_2.png) [@Artemonim](https://forum.cursor.com/u/Artemonim)\ **Post date:** [July 9, 2025, 10:01am UTC](https://forum.cursor.com/t/agent-docstrings-serve-agents-a-clear-table-of-contents-right-in-your-code/110826/5 "2025-07-09T10:01:54Z")
> [@Artemonim](#): > > I hope I’ve finally cleaned up all the bugs… 🫠 [nope](https://github.com/Artemonim/AgentDocstrings/issues/22) T\_T ---
**Author:** ![Artemonim](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/artemonim/32/56113_2.png) [@Artemonim](https://forum.cursor.com/u/Artemonim)\ **Post date:** [July 26, 2025, 7:29am UTC](https://forum.cursor.com/t/agent-docstrings-serve-agents-a-clear-table-of-contents-right-in-your-code/110826/6 "2025-07-26T07:29:04Z")
Finally released another ecosystem tool: > [@Agent Enforcer Open Beta: Enforce the Quality of Agent Code](https://forum.cursor.com/t/agent-enforcer-open-beta-enforce-the-quality-of-agent-code/122880): > > Hi! I’ve finally managed to synthesize my scattered scripts into Agent Enforcer. Theoretically, it’s supposed to support Python, Kotlin, C#, and JS/TS. But personally, I’ve only tested it on Python. By the way, I have no idea what range of language versions it supports smiling_face_with_sunglasses I also haven’t had the time to double-check all the rules the Enforcer uses for the supported languages. But by default, it should check the code pretty strictly. Despite the high test coverage, …