YurunChen/repo-docs-skills

| --- | --- | --- | | Normalized record | Stable handoff after input parsing. Later code reads this record, not the raw input. Often confused with the original file, request, or row. | — | | OTLP | Trace export format this repo uses. It is 

| Normalized record | Stable handoff after input parsing. Later code reads this record, not the raw input. Often confused with the original file, request, or row. | — | | OTLP | Trace export format this repo uses. It is different from Prome

| OTLP | Trace export format this repo uses. It is different from Prometheus scraping. | Inferred — [OTLP spec](https://opentelemetry.io/docs/specs/otlp/) |

</p>   <p>     <a href="http://xhslink.com/o/9M27ebWDaD3">

<p>     <a href="http://xhslink.com/o/9M27ebWDaD3">       <img alt="RedNote(小红书)" src="https://img.shields.io/badge/RedNote(%E5%B0%8F%E7%BA%A2%E4%B9%A6)-Post-ff2442?style=for-the-badge&amp;labelColor=555" />

<a href="http://xhslink.com/o/9M27ebWDaD3">

<img alt="RedNote(小红书)" src="https://img.shields.io/badge/RedNote(%E5%B0%8F%E7%BA%A2%E4%B9%A6)-Post-ff2442?style=for-the-badge&amp;labelColor=555" />

</a>     <a href="https://yurunchen.github.io/repo-docs-skills/">       <img alt="Project homepage" src="https://img.shields.io/badge/Project%20Page-Online-27a36f?style=for-the-badge&amp;labelColor=555" />

<a href="https://yurunchen.github.io/repo-docs-skills/">

<img alt="Project homepage" src="https://img.shields.io/badge/Project%20Page-Online-27a36f?style=for-the-badge&amp;labelColor=555" />

<a href="README_CN.md">Chinese README</a> |   <a href="SKILL.md">Skill contract</a> |   <a href="https://yurunchen.github.io/repo-docs-skills/">Project homepage</a> |

<a href="SKILL.md">Skill contract</a> |   <a href="https://yurunchen.github.io/repo-docs-skills/">Project homepage</a> |   <a href="#install-in-30-seconds">Install</a>

<a href="https://yurunchen.github.io/repo-docs-skills/">Project homepage</a> |

AI coding is no longer a niche workflow. Two 2026 open-source studies make the scale visible: [AIDev](https://arxiv.org/abs/2602.09185) reports 932,791 agent-authored pull requests across 116,211 GitHub repositories, while a

scale visible: [AIDev](https://arxiv.org/abs/2602.09185) reports 932,791 agent-authored pull requests across 116,211 GitHub repositories, while a [multi-method census of 180 million repositories](https://arxiv.org/abs/2606.24429)

agent-authored pull requests across 116,211 GitHub repositories, while a [multi-method census of 180 million repositories](https://arxiv.org/abs/2606.24429) shows that many agent traces are missed by single-signal detection.

[multi-method census of 180 million repositories](https://arxiv.org/abs/2606.24429)

Install the repo-docs skill from this project: https://github.com/YurunChen/repo-docs-skills

https://github.com/YurunChen/repo-docs-skills

curl -fsSL https://github.com/YurunChen/repo-docs-skills/raw/main/install.sh | bash

curl -fsSL https://github.com/YurunChen/repo-docs-skills/raw/main/install.sh | bash

irm https://github.com/YurunChen/repo-docs-skills/raw/main/install.ps1 | iex

Repo Docs Skills is developed by the [AI4GC Lab](https://ai4gc.org/) at Zhejiang University.

- [codebase-to-course](https://github.com/zarazhangrui/codebase-to-course)

- [neat-freak](https://github.com/KKKKhazix/khazix-skills)

<br />   <br />   <img src="https://visitor-badge.laobi.icu/badge?page_id=YurunChen.Repo-Docs" alt="visitors" />

<br />   <img src="https://visitor-badge.laobi.icu/badge?page_id=YurunChen.Repo-Docs" alt="visitors" /> </div>

<img src="https://visitor-badge.laobi.icu/badge?page_id=YurunChen.Repo-Docs" alt="visitors" />

</p>   <p>     <a href="http://xhslink.com/o/9M27ebWDaD3">

<p>     <a href="http://xhslink.com/o/9M27ebWDaD3">       <img alt="RedNote(小红书)" src="https://img.shields.io/badge/RedNote(%E5%B0%8F%E7%BA%A2%E4%B9%A6)-Post-ff2442?style=for-the-badge&amp;labelColor=555" />

<a href="http://xhslink.com/o/9M27ebWDaD3">

<img alt="RedNote(小红书)" src="https://img.shields.io/badge/RedNote(%E5%B0%8F%E7%BA%A2%E4%B9%A6)-Post-ff2442?style=for-the-badge&amp;labelColor=555" />

</a>     <a href="https://yurunchen.github.io/repo-docs-skills/">       <img alt="项目主页" src="https://img.shields.io/badge/Project%20Page-Online-27a36f?style=for-the-badge&amp;labelColor=555" />

<a href="https://yurunchen.github.io/repo-docs-skills/">

<img alt="项目主页" src="https://img.shields.io/badge/Project%20Page-Online-27a36f?style=for-the-badge&amp;labelColor=555" />

<a href="README.md">English README</a> |   <a href="SKILL.md">Skill contract</a> |   <a href="https://yurunchen.github.io/repo-docs-skills/">项目主页</a> |

<a href="SKILL.md">Skill contract</a> |   <a href="https://yurunchen.github.io/repo-docs-skills/">项目主页</a> |   <a href="#30-秒安装">安装</a>

<a href="https://yurunchen.github.io/repo-docs-skills/">项目主页</a> |

AI coding 已经不是小众玩法。两篇 2026 年的开源研究让这个趋势变得可量化： [AIDev](https://arxiv.org/abs/2602.09185) 报告了 932,791 个 agent-authored pull requests，覆盖 116,211 个 GitHub 仓库；另一篇

[AIDev](https://arxiv.org/abs/2602.09185) 报告了 932,791 个 agent-authored pull requests，覆盖 116,211 个 GitHub 仓库；另一篇 [180M 仓库的多方法普查](https://arxiv.org/abs/2606.24429) 则显示，很多

pull requests，覆盖 116,211 个 GitHub 仓库；另一篇 [180M 仓库的多方法普查](https://arxiv.org/abs/2606.24429) 则显示，很多 agent 痕迹会被单一信号漏掉。

[180M 仓库的多方法普查](https://arxiv.org/abs/2606.24429) 则显示，很多

Install the repo-docs skill from this project: https://github.com/YurunChen/repo-docs-skills

https://github.com/YurunChen/repo-docs-skills

curl -fsSL https://github.com/YurunChen/repo-docs-skills/raw/main/install.sh | bash

curl -fsSL https://github.com/YurunChen/repo-docs-skills/raw/main/install.sh | bash

irm https://github.com/YurunChen/repo-docs-skills/raw/main/install.ps1 | iex

Repo Docs Skills 由浙江大学 [AI4GC Lab](https://ai4gc.org/) 开发。

- [codebase-to-course](https://github.com/zarazhangrui/codebase-to-course)

- [neat-freak](https://github.com/KKKKhazix/khazix-skills)

<br />   <br />   <img src="https://visitor-badge.laobi.icu/badge?page_id=YurunChen.Repo-Docs" alt="visitors" />

<br />   <img src="https://visitor-badge.laobi.icu/badge?page_id=YurunChen.Repo-Docs" alt="visitors" /> </div>

<img src="https://visitor-badge.laobi.icu/badge?page_id=YurunChen.Repo-Docs" alt="visitors" />

/*!  * GSAP 3.13.0  * https://gsap.com

Executable file

* GSAP 3.13.0  * https://gsap.com  *

* https://gsap.com

*   * @license Copyright 2025, GreenSock. All rights reserved.  * Subject to the terms at https://gsap.com/standard-license.

* @license Copyright 2025, GreenSock. All rights reserved.  * Subject to the terms at https://gsap.com/standard-license.  * @author: Jack Doyle, jack@greensock.com

* Subject to the terms at https://gsap.com/standard-license.

!function(t,e){"object"==typeof exports&&"undefined"!=typeof module?e(exports):"function"==typeof define&&define.amd?define(["exports"],e):e((t=t||self).window=t.window||{})}(this,function(e){"use strict";function _inheritsLoose(t,e){t.prot

background:     linear-gradient(180deg, rgb(255 255 255 / 0.04), transparent 24rem),     url("data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' width='160' height='160'%3E%3Cfilter id='n'%3E%3CfeTurbulence type='fractalNoise' bas

linear-gradient(180deg, rgb(255 255 255 / 0.04), transparent 24rem),     url("data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' width='160' height='160'%3E%3Cfilter id='n'%3E%3CfeTurbulence type='fractalNoise' baseFrequency='0.85

url("data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' width='160' height='160'%3E%3Cfilter id='n'%3E%3CfeTurbulence type='fractalNoise' baseFrequency='0.85' numOctaves='2'/%3E%3C/filter%3E%3Crect width='100%25' height='100%25' f

Executable file

const INSTALL_TEXT = {     en: `Install the repo-docs skill from this project: https://github.com/YurunChen/repo-docs-skills

https://github.com/YurunChen/repo-docs-skills

in my agent skill directory.`,     zh: `从这个项目安装 repo-docs skill: https://github.com/YurunChen/repo-docs-skills

https://github.com/YurunChen/repo-docs-skills

"[data-copy-install]": "Copy",         ".hero-actions .button-primary": "See generated docs",         ".hero-actions a[href='https://github.com/YurunChen/repo-docs-skills']": "GitHub",

".hero-actions .button-primary": "See generated docs",         ".hero-actions a[href='https://github.com/YurunChen/repo-docs-skills']": "GitHub",         ".hero-actions a[href='#contribute']": "Contribute",

".hero-actions a[href='https://github.com/YurunChen/repo-docs-skills']": "GitHub",

"[data-copy-install]": "复制",         ".hero-actions .button-primary": "查看生成文档",         ".hero-actions a[href='https://github.com/YurunChen/repo-docs-skills']": "GitHub",

".hero-actions .button-primary": "查看生成文档",         ".hero-actions a[href='https://github.com/YurunChen/repo-docs-skills']": "GitHub",         ".hero-actions a[href='#contribute']": "参与贡献",

".hero-actions a[href='https://github.com/YurunChen/repo-docs-skills']": "GitHub",

try {       const response = await fetch(trigger.dataset.file);

const response = await fetch(trigger.dataset.file);

This case applies the `repo-docs` build shape to [Aider-AI/aider](https://github.com/Aider-AI/aider), a popular terminal pair-programming assistant.

| Path | Responsibility | Key code | Connection to the main path | | --- | --- | --- | --- | | `aider/` | Turns terminal, config, model, file, and git state into one configured coding session. | [`main.py`](https://github.com/Aider-AI/aider

| --- | --- | --- | --- | | `aider/` | Turns terminal, config, model, file, and git state into one configured coding session. | [`main.py`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/main.py#L451) 

| `aider/` | Turns terminal, config, model, file, and git state into one configured coding session. | [`main.py`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/main.py#L451) | Owns startup and the pol

| `aider/coders/` | Runs the model turn and constrains model text into permitted, checkable file edits. | [`base_coder.py`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/coders/base_coder.py#L124), [`

| `tests/basic/` | Verifies edit parsing, dry runs, dirty-file protection, and commit scope. | [`test_editblock.py`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/tests/basic/test_editblock.py), [`test_code

The main source anchor is [`main()` from git/config discovery through parsing](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/main.py#L451), followed by [`Coder.create()` receiving the session policy](

`Coder.create()` turns the requested or model-default edit format into a concrete class. If the edit format changes while switching coders, it can summarize prior chat history so the old edit-format examples do not confuse the new coder. Th

| Pass | Purpose | Inspected evidence | What changed in the model | | --- | --- | --- | --- | | Pass 1 | Map the main edit path from terminal input to changed files. | [`pyproject.toml`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f97

| --- | --- | --- | --- | | Pass 1 | Map the main edit path from terminal input to changed files. | [`pyproject.toml`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/pyproject.toml), [`aider/__main__.py`](ht

| Pass 1 | Map the main edit path from terminal input to changed files. | [`pyproject.toml`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/pyproject.toml), [`aider/__main__.py`](https://github.com/Aider-AI/

| Pass 2 | Challenge the path for boundaries, failure modes, and tests. | [`tests/basic/test_main.py`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/tests/basic/test_main.py), [`tests/basic/test_coder.py`](

| --- | --- | | Scenario | A developer starts `aider` in a git repo, gives a coding request, and the default diff-style coder applies a SEARCH/REPLACE edit to a file in the chat. | | Input | The observable input is CLI arguments and a user 

| Scenario | A developer starts `aider` in a git repo, gives a coding request, and the default diff-style coder applies a SEARCH/REPLACE edit to a file in the chat. | | Input | The observable input is CLI arguments and a user message. The p

| Input | The observable input is CLI arguments and a user message. The package exposes the `aider` command through `aider.main:main` in [`pyproject.toml`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/pypr

| Success / output | A file is changed, and when git auto-commit is enabled the edited file can be committed with an Aider-generated message. Tests assert changed file content and commit counts in [`tests/basic/test_coder.py`](https://githu

| Claim | Evidence | Confidence | Caveat | Used by | | --- | --- | --- | --- | --- | | The installed CLI command enters `aider.main:main`. | [`pyproject.toml`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/

| --- | --- | --- | --- | --- | | The installed CLI command enters `aider.main:main`. | [`pyproject.toml`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/pyproject.toml) declares `aider = "aider.main:main"`,

| The installed CLI command enters `aider.main:main`. | [`pyproject.toml`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/pyproject.toml) declares `aider = "aider.main:main"`, and [`aider/__main__.py`](https

| `main()` resolves git/config context before creating a coder. | [`aider/main.py`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/main.py) starts with git root detection, config search, parser constru

| `main()` assembles the objects that drive an edit session and passes them into `Coder.create()`. | [`aider/main.py`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/main.py) creates `Commands`, `ChatS

| `Coder.create()` chooses the concrete coder class by edit format. | [`aider/coders/base_coder.py`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/coders/base_coder.py) resolves `edit_format` from the

| The default diff-style coder parses SEARCH/REPLACE blocks from model output. | [`aider/coders/editblock_coder.py`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/coders/editblock_coder.py) has `EditB

| `send_message()` owns the model-call loop, retries, output-limit handling, response capture, and post-response edit phase. | [`aider/coders/base_coder.py`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/ai

| File writes are gated by a dry-run parse before permission checks and final writes. | [`aider/coders/base_coder.py`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/coders/base_coder.py) calls `get_ed

| Aider asks before creating or editing files outside the current chat set, and skips gitignored files. | [`aider/coders/base_coder.py`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/coders/base_coder

| Dirty files targeted by model edits can be committed before the edit, preserving an undoable baseline. | [`aider/coders/base_coder.py`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/coders/base_code

| Aider commits only edited files for an Aider edit, not unrelated dirty files. | [`tests/basic/test_coder.py`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/tests/basic/test_coder.py) keeps `file1.txt` dir

| Non-matching SEARCH blocks become reflected feedback instead of silent success. | [`aider/coders/editblock_coder.py`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/coders/editblock_coder.py) builds 

| After edits, Aider may auto-commit, run lint, reflect lint errors, run shell commands, and run tests. | [`aider/coders/base_coder.py`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/coders/base_coder

| `--apply` reuses the edit application path without sending a model request. | [`aider/main.py`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/main.py) reads an apply file into `coder.partial_respons

Current verification caveat: the focused pytest command was attempted on 2026-06-29 13:30 CST, but this Python environment is missing `oslex`, so pytest failed during import before executing the selected tests. `oslex==0.1.3` is listed in [

The source anchor is [`main()` setting git/config/parser state](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/main.py#L451). The config behavior is tested in [`test_main_with_git_config_yml`](https://

The source anchor is [`main()` calling `Coder.create()` with session policy](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/main.py#L972). The concrete class is chosen by [`Coder.create()` matching `ed

The source anchor is [`send_message()` formatting, sending, and capturing the response](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/coders/base_coder.py#L1419). The provider call itself is deliberat

The source anchors are [`EditBlockCoder.get_edits()`](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/coders/editblock_coder.py#L21), [`EditBlockCoder.apply_edits()`](https://github.com/Aider-AI/aider/b

The source anchor is [`allowed_to_edit()` checking in-chat, gitignore, new-file, and out-of-chat cases](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/coders/base_coder.py#L2191). The new-file path is 

The source anchors are [`check_for_dirty_commit()` marking dirty files](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/coders/base_coder.py#L2175), [`prepare_to_edit()` calling `dirty_commit()`](https:

The source anchor is [`send_message()` handling auto-commit, lint, shell output, and tests after edits](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/coders/base_coder.py#L1585). [`test_only_commit_gp

The source anchors are [`EditBlockCoder.apply_edits()` constructing the no-match message](https://github.com/Aider-AI/aider/blob/5dc9490bb35f9729ef2c95d00a19ccd30c26339c/aider/coders/editblock_coder.py#L82) and [`apply_updates()` reflecting

| Path | Responsibility | Key code | Connection to the main path | | --- | --- | --- | --- | | `app/components/chat/` | Builds the browser-side request from the user message, workbench state, and selected provider/model. | [`Chat.client.tsx

| --- | --- | --- | --- | | `app/components/chat/` | Builds the browser-side request from the user message, workbench state, and selected provider/model. | [`Chat.client.tsx`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a69639

| `app/components/chat/` | Builds the browser-side request from the user message, workbench state, and selected provider/model. | [`Chat.client.tsx`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/

| `app/routes/` | Orchestrates the server request, context work, response continuation, annotations, and user-facing stream errors. | [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/

| Field | Meaning | Source | | --- | --- | --- | | `apiKeys` | Provider API keys from client state. | [`Chat.client.tsx`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/components/chat/Chat.cli

| --- | --- | --- | | `apiKeys` | Provider API keys from client state. | [`Chat.client.tsx`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/components/chat/Chat.client.tsx#L136) | | `files` | C

| `apiKeys` | Provider API keys from client state. | [`Chat.client.tsx`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/components/chat/Chat.client.tsx#L136) |

| `files` | Current workbench file map, used for context selection and locked-file prompts. | [`Chat.client.tsx`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/components/chat/Chat.client.tsx#

| `promptId` | Prompt-library selection. | [`Chat.client.tsx`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/components/chat/Chat.client.tsx#L139) |

| `contextOptimization` | Enables summary and selected-file context path when files exist. | [`Chat.client.tsx`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/components/chat/Chat.client.tsx#L

| `chatMode` | Chooses build prompt versus discuss prompt. | [`Chat.client.tsx`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/components/chat/Chat.client.tsx#L141) |

| `designScheme` | Optional design prompt input. | [`Chat.client.tsx`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/components/chat/Chat.client.tsx#L142) |

| `supabase` | Optional Supabase connection and credentials. | [`Chat.client.tsx`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/components/chat/Chat.client.tsx#L143) |

| `maxLLMSteps` | Max tool-call steps passed to the AI SDK options. | [`Chat.client.tsx`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/components/chat/Chat.client.tsx#L151) |

The server expects the same fields in `request.json()`. Source: [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/routes/api.chat.ts#L51).

The client writes this format for new chats, template-seeded chats, modified-file chats, and normal appended chats. Source: [`Chat.client.tsx`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/co

The server extracts and removes those tags before sending messages to the model. Source: [`utils.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/lib/.server/llm/utils.ts#L8).

| Cookie | Meaning | Source | | --- | --- | --- | | `apiKeys` | JSON object for provider API keys. | [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/routes/api.chat.ts#L70) |

| --- | --- | --- | | `apiKeys` | JSON object for provider API keys. | [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/routes/api.chat.ts#L70) | | `providers` | JSON object of pr

| `apiKeys` | JSON object for provider API keys. | [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/routes/api.chat.ts#L70) |

| `providers` | JSON object of provider settings. | [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/routes/api.chat.ts#L72) |

The separate model-list route uses shared cookie helpers for the same concepts. Source: [`cookies.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/lib/api/cookies.ts#L13).

| Stream item | When it appears | Source | | --- | --- | --- | | `progress` with `summary` | Before and after chat summary generation. | [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bf

| --- | --- | --- | | `progress` with `summary` | Before and after chat summary generation. | [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/routes/api.chat.ts#L108) | | `chatSu

| `progress` with `summary` | Before and after chat summary generation. | [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/routes/api.chat.ts#L108) |

| `chatSummary` annotation | After summary generation. | [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/routes/api.chat.ts#L145) |

| `progress` with `context` | Before and after context selection. | [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/routes/api.chat.ts#L153) |

| `codeContext` annotation | After selected files are known. | [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/routes/api.chat.ts#L186) |

| `progress` with `response` | Before response generation and after normal completion. | [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/routes/api.chat.ts#L230) |

| `usage` annotation | On normal completion, with cumulative tokens from prework and final response. | [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/routes/api.chat.ts#L231) |

The stream is not only the answer text. It also carries `progress`, `chatSummary`, `codeContext`, and `usage` annotations that let the UI understand what happened around the answer. Source: [progress and annotations in `api.chat.ts`](https:

The route creates an AI SDK data stream, performs pre-response work inside `execute()`, then merges the lower-level AI SDK stream into that data stream. Context optimization is conditional: it only runs when the request has files and `conte

The lower-level `streamText()` call is responsible for prompt construction. In build mode it adds selected code context and an optional summary to the system prompt. It also adds a locked-file warning when any workbench file carries `isLock

The main boundary is selector output. `selectContext()` requires a model-generated XML-like wrapper and throws on invalid format or zero selected files. That keeps context selection explicit, but it also means context optimization can fail 

`extractPropertiesFromMessage()` reads `[Model: ...]` and `[Provider: ...]` from user message content and returns cleaned content. Source: [tag extraction](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b1

`streamText()` then looks up the provider in `PROVIDER_LIST`, checks static models, and asks `LLMManager.getModelListFromProvider()` for dynamic models when static lookup misses. If no provider model list exists, it throws. If the selected 

`LLMManager` registers provider classes exported by the registry, stores them by provider name, and combines static and dynamic models. Dynamic model fetch failures are caught and treated as empty dynamic lists during broad updates, which m

Token parameters are selected after the model is known. Reasoning model names matching `o1`, `o3`, or `gpt-5` use `maxCompletionTokens` and get unsupported sampling options removed. Other models use `maxTokens`. Source: [token parameter sel

| Claim | Evidence | Confidence | Caveat | Used by | | --- | --- | --- | --- | --- | | The project is an AI coding assistant with a Remix web app and many provider integrations. | [`README.md`](https://github.com/stackblitz-labs/bolt.diy/bl

| --- | --- | --- | --- | --- | | The project is an AI coding assistant with a Remix web app and many provider integrations. | [`README.md`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/README.md

| The project is an AI coding assistant with a Remix web app and many provider integrations. | [`README.md`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/README.md), [`package.json`](https://gith

| The client posts chat requests to `/api/chat` with body fields for keys, files, prompt ID, context optimization, chat mode, Supabase state, design scheme, and MCP steps. | [`Chat.client.tsx`](https://github.com/stackblitz-labs/bolt.diy/bl

| Client messages embed `[Model: ...]` and `[Provider: ...]` tags. | New chat and appended chat paths in [`Chat.client.tsx`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/components/chat/Chat.

| The server route parses the same payload fields and reads `apiKeys` / `providers` cookies. | [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/routes/api.chat.ts#L51) | Confirmed

| MCP tool processing happens before context optimization and response generation. | [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/routes/api.chat.ts#L87) and [`api.chat.ts`](h

| Context optimization runs only when files exist and the flag is enabled. | [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/routes/api.chat.ts#L97) and [`api.chat.ts`](https://g

| Summary and context selection contribute to cumulative token usage. | [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/routes/api.chat.ts#L121) and [`api.chat.ts`](https://githu

| `selectContext()` requires `<updateContextBuffer>` output and throws on invalid format or zero selected files. | [`select-context.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/lib/.serv

| Provider/model tags are extracted and cleaned before the model sees content. | [`utils.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/lib/.server/llm/utils.ts#L8) and [`stream-text.ts`](

| `streamText()` resolves provider/model through static and dynamic model lists, with fallback behavior. | [`stream-text.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/lib/.server/llm/stre

| Provider classes are registered through `LLMManager` from the provider registry. | [`manager.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/lib/modules/llm/manager.ts#L33), [`registry.ts

| Build mode adds selected file context, summaries, and locked-file instructions to the system prompt. | [`stream-text.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/lib/.server/llm/stream

| Reasoning models use `maxCompletionTokens`; other models use `maxTokens`, and unsupported options are removed for reasoning models. | [`constants.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bf

| A normal finish writes cumulative usage and response-complete progress. | [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/routes/api.chat.ts#L230) | Confirmed | If the stream f

| Length finishes trigger a synthetic continue prompt until `MAX_RESPONSE_SEGMENTS` is reached. | [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/routes/api.chat.ts#L252), [`cons

| Common provider/API errors are converted into custom user-facing messages. | [`api.chat.ts`](https://github.com/stackblitz-labs/bolt.diy/blob/2e254ac19a696394030601bc602f54945b12bfc4/app/routes/api.chat.ts#L349) | Confirmed | Error matchi

The first handoff happens before the server sees anything. The chat UI configures `useChat` to post to `/api/chat` and includes the runtime body: API keys, current files, prompt ID, context-optimization setting, chat mode, design scheme, Su

When the user sends a new message, the client prepends `[Model: ...]` and `[Provider: ...]` to the text. On a new chat it calls `reload()` after `setMessages()`. On an existing chat it calls `append()`, and if the workbench has modified fil

The server route starts by parsing the JSON payload and reading provider settings from cookies. The fields it expects are `messages`, `files`, `promptId`, `contextOptimization`, `chatMode`, optional `designScheme`, optional Supabase credent

Once streaming begins, the route starts stream recovery monitoring, asks the MCP service to process any tool invocations already present in the messages, and records where it may slice older messages if a summary is later available. Source:

The visible output starts as server-sent stream data. The route writes progress annotations such as "Analysing Request", "Determining Files to Read", and "Generating Response", then returns a `text/event-stream` response. The exact stream f

If the request includes files and context optimization is enabled, the server does not send the full project blindly. It first creates a chat summary, writes a `chatSummary` annotation, then asks `selectContext()` to pick relevant files fro

The selector has a strict format boundary. It shows the model available file paths, current context buffer, and the last user question, then requires a response wrapped in `<updateContextBuffer>`. It accepts `<includeFile path="..."/>` and 

Before calling the AI SDK, `streamText()` scans user messages for the provider/model tags and removes them from the content. It then finds the provider in the registered provider list, tries static models first, asks the provider for dynami

The model call also builds the correct prompt for the mode. In build mode, selected context files are wrapped as a context buffer and any chat summary can shorten the message list. Locked files are appended to the system prompt as explicit 

Finally, reasoning models use `maxCompletionTokens` while regular models use `maxTokens`; unsupported sampling options are stripped for reasoning models before the AI SDK call. Source: [token and option handling](https://github.com/stackbli

The route calls `streamText()` and merges the result into the data stream. As chunks arrive, stream recovery activity is updated; if an error chunk appears, the route logs it and stops recovery. Source: [the main stream merge](https://githu

A normal finish writes a `usage` annotation with cumulative prompt, completion, and total tokens, then emits a completed response progress update. If the AI SDK reports `finishReason === 'length'`, the route appends the partial assistant co

User-facing failure messages are shaped in one place. Model-not-found, invalid JSON responses, missing API keys, token limits, rate limits, and network/timeout errors are mapped to custom error text before the stream reaches the client. Sou

| Field | Meaning in this path | Source | | --- | --- | --- | | `language` | Optional VS Code-style language identifier; falls back to `unknown` if absent. | [`CompletionRequest`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72

| --- | --- | --- | | `language` | Optional VS Code-style language identifier; falls back to `unknown` if absent. | [`CompletionRequest`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/servic

| `language` | Optional VS Code-style language identifier; falls back to `unknown` if absent. | [`CompletionRequest`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs#L41

| `segments.prefix` | Text before the cursor. This is the main prompt input for segment-based completion. | [`Segments`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs#

| `segments.suffix` | Optional text after the cursor. Empty or missing suffix becomes a newline during prompt templating. | [`get_default_suffix`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/s

| `segments.filepath` | Optional path for the edited file; used when building indexed-code search queries. | [`Segments`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs

| `segments.git_url` | Optional repository identity; indexed search requires it and an allowed repository match. | [`PromptBuilder::collect`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/se

| `segments.declarations` | Editor-provided declaration snippets with highest snippet priority. | [`extract_snippets_from_segments`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/co

| `segments.relevant_snippets_from_changed_files` | Editor-provided snippets from recently changed files, used after declarations. | [`extract_snippets_from_segments`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e

| `segments.relevant_snippets_from_recently_opened_files` | Editor-provided snippets from recently opened files, used after changed-file snippets. | [`extract_snippets_from_segments`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b983

| `debug_options.raw_prompt` | Direct model prompt; when present, segment-based prompt construction is skipped. | [`raw_prompt`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/comple

| `debug_options.return_snippets` | Includes collected snippets in `debug_data`. | [`DebugOptions`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs#L110-L128), [`debug_d

| `debug_options.return_prompt` | Includes final prompt in `debug_data`. | [`DebugOptions`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs#L110-L128), [`debug_data`](ht

| `debug_options.disable_retrieval_augmented_code_completion` | Skips snippet collection. | [`build_snippets`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs#L315-L329)

| `temperature` | Optional sampling temperature passed into generation options. | [`text_generation_options`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs#L331-L356) 

| `seed` | Optional generation seed passed into generation options. | [`text_generation_options`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs#L331-L356) |

| `mode` | Defaults to `standard`; `next_edit_suggestion` takes a separate prompt path requiring `segments.edit_history`. | [`default_standard_mode`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabb

| Field | Meaning | Source | | --- | --- | --- | | `id` | Generated completion id with `cmpl-` prefix. | [`generate`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs#L35

| --- | --- | --- | | `id` | Generated completion id with `cmpl-` prefix. | [`generate`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs#L358-L365) | | `choices[0].text`

| `id` | Generated completion id with `cmpl-` prefix. | [`generate`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs#L358-L365) |

| `choices[0].text` | Generated model text. | [`Choice`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs#L223-L233), [`generate`](https://github.com/TabbyML/tabby/blob/e

| `debug_data.snippets` | Present only when requested by debug options. | [`DebugData`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs#L274-L280), [`generate`](https://

| `debug_data.prompt` | Present only when requested by debug options. | [`DebugData`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs#L274-L280), [`generate`](https://gi

| `mode` | `standard` for normal completions, `next_edit_suggestion` for next-edit mode. | [`CompletionResponse`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs#L242-L2

| Item | Value or behavior | Source | | --- | --- | --- | | Default host | `0.0.0.0` | [`ServeArgs`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/serve.rs#L93-L97) |

| --- | --- | --- | | Default host | `0.0.0.0` | [`ServeArgs`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/serve.rs#L93-L97) | | Default port | `8080` | [`ServeArgs`](https://github.com/Ta

| Default host | `0.0.0.0` | [`ServeArgs`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/serve.rs#L93-L97) |

| Default port | `8080` | [`ServeArgs`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/serve.rs#L93-L97) |

| Default device | `cpu` | [`ServeArgs`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/serve.rs#L99-L101) |

| Default completion timeout | `30` seconds | [`ServerConfig`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby-common/src/config.rs#L233-L245) |

| Default completion max input length | `1536` characters | [`CompletionConfig`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby-common/src/config.rs#L386-L418) |

| Default max decoding tokens | `64` | [`CompletionConfig`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby-common/src/config.rs#L386-L418) |

| Completion route absent model | `/v1/completions` returns `501 Not Implemented` when no completion service exists. | [`api_router`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/serve.rs#L

| Service error at route | Route logs the error and returns `400 Bad Request`. | [`crates/tabby/src/routes/completions.rs`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/routes/completions.r

curl -X 'POST' \

'http://localhost:8080/v1/completions' \

The route/service split is visible in [`crates/tabby/src/routes/completions.rs`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/routes/completions.rs#L26-L48) and [`CompletionService::generat

Source: [`PromptBuilder::collect`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion/completion_prompt.rs#L40-L86), [`extract_snippets_from_segments`](https://github.com/Tabb

| Claim | Evidence | Confidence | Caveat | Used by | | --- | --- | --- | --- | --- | | Tabby is a self-hosted AI coding assistant with an OpenAPI-compatible server surface. | [`README.md`](https://github.com/TabbyML/tabby/blob/e8608d6d8f401

| --- | --- | --- | --- | --- | | Tabby is a self-hosted AI coding assistant with an OpenAPI-compatible server surface. | [`README.md`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/README.md) and [`serve.rs

| Tabby is a self-hosted AI coding assistant with an OpenAPI-compatible server surface. | [`README.md`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/README.md) and [`serve.rs` OpenAPI paths](https://github.

| The scoped run starts with `Serve` and dispatches into `serve::main`. | [`main.rs`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/main.rs#L53-L73). | Confirmed | `download` subcommand is n

| CLI `--model` can override configured completion model and create local model config. | [`ServeArgs`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/serve.rs#L83-L115), [`merge_args`](https

| `/v1/completions` is installed only when a completion service exists; otherwise it returns `501`. | [`api_router`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/serve.rs#L289-L310). | Conf

| Completion requests are wrapped in a configurable timeout defaulting to 30 seconds. | [`api_router` timeout layer](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/serve.rs#L289-L297), [`Serv

| The route enriches the request with user/user-agent context and maps service errors to `400`. | [`crates/tabby/src/routes/completions.rs`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/rou

| A standard completion needs either `debug_options.raw_prompt` or `segments`. | [`raw_prompt`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs#L90-L95), [`generate`](ht

| Segment-based requests can collect snippets unless retrieval augmentation is disabled. | [`build_snippets`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs#L315-L329).

| Editor-provided snippets are prioritized before indexed-code search. | [`extract_snippets_from_segments`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion/completion_promp

| Indexed search is skipped when code search, `git_url`, or allowed repository match is absent. | [`PromptBuilder::collect`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion

| Code-search `NotReady` and other search errors do not fail completion. | [`collect_snippets`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion/completion_prompt.rs#L224-L2

| Snippets are prepended as comments when the language has a line-comment marker. | [`build_prefix`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion/completion_prompt.rs#L1

| CRLF input is normalized before inference and restored in generated output. | [`contains_crlf`, `override_prompt`, `override_generated_text`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/

| The service logs `Event::Completion` and returns a one-choice response. | [`generate`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs#L407-L438), [`CompletionResponse

| Golden tests start a real server and post completion JSON to `/v1/completions`. | [`goldentests.rs`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/tests/goldentests.rs#L34-L159). | Confirmed |

Source: [`main.rs` dispatches `Serve`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/main.rs#L53-L73), [`serve.rs` builds services and routes](https://github.com/TabbyML/tabby/blob/e8608d6d8

Source: [`crates/tabby/src/routes/completions.rs`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/routes/completions.rs#L26-L48).

Source: [`CompletionRequest` fields](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs#L41-L65), [`CompletionService::generate`](https://github.com/TabbyML/tabby/blob/e860

Source: [`build_snippets` handles the disable flag](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs#L315-L329), [`PromptBuilder::collect` gates indexed retrieval](https:

Source: [`rewrite_with_snippets` and `build_prefix`](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion/completion_prompt.rs#L100-L143), [`build_prompt` and default suffix beh

Source: [`CompletionService::generate` calls the engine, logs, and returns](https://github.com/TabbyML/tabby/blob/e8608d6d8f4016b9836a72037f72630d7e993468/crates/tabby/src/services/completion.rs#L407-L438), [`CompletionResponse` shape](http

<link rel="icon" href="assets/logo.png" />     <link rel="preload" as="image" href="assets/repo-docs-pixel-kingdom.png" fetchpriority="high" />     <link rel="preconnect" href="https://fonts.googleapis.com" />

<link rel="preload" as="image" href="assets/repo-docs-pixel-kingdom.png" fetchpriority="high" />     <link rel="preconnect" href="https://fonts.googleapis.com" />     <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin />

<link rel="preconnect" href="https://fonts.googleapis.com" />

<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin />

<link       href="https://fonts.googleapis.com/css2?family=Spectral:ital,wght@0,500;0,600;0,700;1,500&family=Source+Sans+3:ital,wght@0,400;0,500;0,600;0,700;1,400&display=swap"       rel="stylesheet"

href="https://fonts.googleapis.com/css2?family=Spectral:ital,wght@0,500;0,600;0,700;1,500&family=Source+Sans+3:ital,wght@0,400;0,500;0,600;0,700;1,400&display=swap"

</div>               <pre><code id="install-snippet-code" data-install-snippet>Install the repo-docs skill from this project: https://github.com/YurunChen/repo-docs-skills

https://github.com/YurunChen/repo-docs-skills

<div class="hero-actions" aria-label="Primary actions">               <a class="button button-primary" href="#cases">See generated docs</a>               <a class="button button-secondary" href="https://github.com/YurunChen/repo-docs-skills

<a class="button button-primary" href="#cases">See generated docs</a>               <a class="button button-secondary" href="https://github.com/YurunChen/repo-docs-skills">GitHub</a>               <a class="button button-secondary" href="#c

<a class="button button-secondary" href="https://github.com/YurunChen/repo-docs-skills">GitHub</a>

</div>           <div class="contribute-actions">             <a class="contribute-action contribute-action--primary" href="https://github.com/YurunChen/repo-docs-skills/issues">

<div class="contribute-actions">             <a class="contribute-action contribute-action--primary" href="https://github.com/YurunChen/repo-docs-skills/issues">               <span>GitHub Issues</span>

<a class="contribute-action contribute-action--primary" href="https://github.com/YurunChen/repo-docs-skills/issues">

<strong>Open issue</strong>             </a>             <a class="contribute-action" href="https://github.com/YurunChen/repo-docs-skills/pulls">

</a>             <a class="contribute-action" href="https://github.com/YurunChen/repo-docs-skills/pulls">               <span>Pull Requests</span>

<a class="contribute-action" href="https://github.com/YurunChen/repo-docs-skills/pulls">               <span>Pull Requests</span>               <strong>Send PR</strong>

#!/usr/bin/env python3

#!/usr/bin/env python3

| --- | --- | --- | | Normalized record | Stable handoff after input parsing. Later code reads this record, not the raw input. Often confused with the original file, request, or row. | — | | OTLP | Trace export format this repo uses. It is 

| --- | --- | --- | | Normalized record | Stable handoff after input parsing. Later code reads this record, not the raw input. Often confused with the original file, request, or row. | — | | OTLP | Trace export format this repo uses. It is 

| Normalized record | Stable handoff after input parsing. Later code reads this record, not the raw input. Often confused with the original file, request, or row. | — | | OTLP | Trace export format this repo uses. It is different from Prome

| Normalized record | Stable handoff after input parsing. Later code reads this record, not the raw input. Often confused with the original file, request, or row. | — | | OTLP | Trace export format this repo uses. It is different from Prome

| OTLP | Trace export format this repo uses. It is different from Prometheus scraping. | Inferred — [OTLP spec](https://opentelemetry.io/docs/specs/otlp/) |

| OTLP | Trace export format this repo uses. It is different from Prometheus scraping. | Inferred — [OTLP spec](https://opentelemetry.io/docs/specs/otlp/) |

#!/usr/bin/env python3

#!/usr/bin/env python3

#!/usr/bin/env python3