From 3a9f101d22bdd8b23b5a22def931874e4a14623c Mon Sep 17 00:00:00 2001 From: Alex Romano Date: Fri, 23 May 2025 13:35:30 -0700 Subject: [PATCH 01/11] update agentic chat docs for GA --- docs/cody/capabilities/agentic-chat.mdx | 37 ++++++++++++------------- 1 file changed, 17 insertions(+), 20 deletions(-) diff --git a/docs/cody/capabilities/agentic-chat.mdx b/docs/cody/capabilities/agentic-chat.mdx index 610d739fe..67693f6e4 100644 --- a/docs/cody/capabilities/agentic-chat.mdx +++ b/docs/cody/capabilities/agentic-chat.mdx @@ -1,25 +1,23 @@ -# Agentic chat +# Agentic context fetching -

Learn about the agentic chat experience, an exclusive chat-based AI agent with enhanced capabilities.

+

Learn about agentic context fetching, a mini-agent that uses search and tools to retrieve context.

-Agentic chat (available in version 6.0) is currently in the Experimental stage for Enterprise and is supported on VS Code, JetBrains, Visual Studio editor extensions and Web. Usage may be limited at this stage. +Cody's agentic context fetching experience can evaluate context and fetch any additional context (MCP, OpenCtx, terminal, etc.) by providing enhanced, context-aware chat capabilities. It extends Cody's functionality by proactively understanding your coding environment and gathering relevant information based on your requests before responding. These features help you get noticeably higher-quality responses. -Cody's agentic chat experience is an AI agent that can evaluate context and fetch any additional context (OpenCtx, terminal, etc.) by providing enhanced, context-aware chat capabilities. It extends Cody's functionality by proactively understanding your coding environment and gathering relevant information based on your requests before responding. These features help you get noticeably higher-quality responses. - -This agentic chat experience aims to reduce the learning curve associated with traditional coding assistants by minimizing users' need to provide context manually. It achieves this through agentic context retrieval, where the AI autonomously gathers and analyzes context before generating a response. +This experience aims to reduce the learning curve associated with traditional coding assistants by minimizing users' need to provide context manually. It achieves this through agentic context retrieval, where the AI autonomously gathers and analyzes context before generating a response. ## Capabilities of agentic chat -The agentic chat experience leverages several key capabilities, including: +The agentic context fetching experience leverages several key capabilities, including: - **Proactive context gathering**: Automatically gathers relevant context from your codebase, project structure, and current task - **Agentic context reflection**: Review the gathered context to ensure it is comprehensive and relevant to your query - **Iterative context improvement**: Performs multiple review loops to refine the context and ensure a thorough understanding - **Enhanced response accuracy**: Leverages comprehensive context to provide more accurate and relevant responses, reducing the risk of hallucinations -## What can agentic chat do? +## What can agentic context fetching do? -Agentic chat can help you with the following: +Agentic context fetching can help you with the following: ### Tool Usage @@ -29,6 +27,7 @@ It has access to a suite of tools for retrieving relevant context. These tools i - **Codebase File**: Retrieves the full content from a file in your codebase - **Terminal**: Executes shell commands in your terminal - **Web Browser**: Searches the web for live context +- **MCP**: (Configure MCP and add servers)[] to fetch external context - **OpenCtx**: Any OpenCtx providers could be used by the agent It integrates seamlessly with external services, such as web content retrieval and issue tracking systems, using OpenCtx providers. To learn more, [read the OpenCtx docs](/cody/capabilities/openctx). @@ -37,9 +36,9 @@ It integrates seamlessly with external services, such as web content retrieval a ## Terminal access -Agentic chat can use the CLI Tool to request the execution of shell commands to gather context from your terminal. Its ability to execute terminal commands enhances its context-gathering capabilities. However, it’s essential to understand that any information accessible via your terminal could potentially be shared with the LLM. It's recommended not to request information that you don't want to share. Here's what you should consider: +Agentic context fetching can use the CLI Tool to request the execution of shell commands to gather context from your terminal. Its ability to execute terminal commands enhances its context-gathering capabilities. However, it’s essential to understand that any information accessible via your terminal could potentially be shared with the LLM. It's recommended not to request information that you don't want to share. Here's what you should consider: -- **Requires user consent**: Agentic chat will pause and ask for permission each time before executing any shell command. +- **Requires user consent**: Agentic context fetching will pause and ask for permission each time before executing any shell command. - **Trusted workspaces only**: Commands can only be executed within trusted workspaces with a valid shell - **Potential data sharing**: Any terminal-accessible information may be shared with the LLM @@ -47,25 +46,23 @@ Commands are generated by the agent/LLM based on your request. Avoid asking it t ## Use cases -Agentic chat can be helpful to assist you with a wide range of tasks, including: +Agentic context fetching can be helpful to assist you with a wide range of tasks, including: - **Improved response quality**: Helps you get better and more accurate responses than other LLMs, making up for the additional processing time for context gathering a non-issue - **Error resolution**: It can automatically identify error sources and suggest fixes by analyzing error logs - **Better unit tests**: Automatically includes imports and other missing contexts to generate better unit tests -## Enable agentic chat - -### Getting agentic chat access for Pro users +## Enable agentic context fetching -Pro users can find the agentic chat option in the LLM selector drop-down. +### Getting agentic context fetching access for Pro users -![agentic chat interface](https://storage.googleapis.com/sourcegraph-assets/Docs/agenti%20-chat.jpg) +Pro users can find the agentic context fetching option in the LLM selector drop-down. -### Getting agentic chat access for Enterprise customers +![agentic context fetching interface](https://storage.googleapis.com/sourcegraph-assets/Docs/agenti%20-chat.jpg) -Enterprise customers must opt-in to access this agentic chat feature (reach out to your account team for access). +### Getting agentic context fetching access for Enterprise customers -For the experimental release, agentic chat is specifically limited to using Claude Haiku for the reflection steps and Claude Sonnet for the final response to provide a good balance between quality and latency. Therefore, your enterprise instance must have access to both Claude Sonnet and Claude Haiku to use agentic chat. We use the latest versions of these models, and can fall back to older versions when necessary. These models may be changed during the experimental phase to optimize for quality and/or latency. +Agentic context fetching uses smaller models from the Gemini, Claude, and GPT families for reflection steps and whichever model you choose from the model selector for the final response. This provide a good balance between quality and latency. If none of the smaller models are available on your instance, we will fall back to the model chosen in the model selector for reflection. We use the latest versions of these models, and can fall back to older versions when necessary. The default models may be changed to optimize for quality and/or latency. Additionally, enterprise users need to upgrade their supported client (VS Code, JetBrains, and Visual Studio) to the latest version of the plugin by enabling the following feature flags on their Sourcegraph Instance: From c2b424ab70a943588259820c92ae73d2644416c0 Mon Sep 17 00:00:00 2001 From: Alex Romano Date: Fri, 23 May 2025 13:43:23 -0700 Subject: [PATCH 02/11] remove ff --- docs/cody/capabilities/agentic-chat.mdx | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/docs/cody/capabilities/agentic-chat.mdx b/docs/cody/capabilities/agentic-chat.mdx index 67693f6e4..3b8c644e0 100644 --- a/docs/cody/capabilities/agentic-chat.mdx +++ b/docs/cody/capabilities/agentic-chat.mdx @@ -64,7 +64,4 @@ Pro users can find the agentic context fetching option in the LLM selector drop- Agentic context fetching uses smaller models from the Gemini, Claude, and GPT families for reflection steps and whichever model you choose from the model selector for the final response. This provide a good balance between quality and latency. If none of the smaller models are available on your instance, we will fall back to the model chosen in the model selector for reflection. We use the latest versions of these models, and can fall back to older versions when necessary. The default models may be changed to optimize for quality and/or latency. -Additionally, enterprise users need to upgrade their supported client (VS Code, JetBrains, and Visual Studio) to the latest version of the plugin by enabling the following feature flags on their Sourcegraph Instance: - -- `agentic-chat-experimental` to get access to the feature -- `agentic-chat-cli-tool-experimental` to allow [terminal access](#terminal-commands) +Terminal access is disabled by default. To enable it, set the `agentic-chat-cli-tool-experimental` feature flag [terminal access](#terminal-commands) From 931e497fb84366fb313fda333474355d773a6e3a Mon Sep 17 00:00:00 2001 From: Maedah Batool Date: Mon, 26 May 2025 20:11:47 -0700 Subject: [PATCH 03/11] Add more docs --- ...-chat.mdx => agentic-context-fetching.mdx} | 20 +++++++------------ docs/cody/clients/feature-reference.mdx | 11 +++++----- src/data/navigation.ts | 2 +- src/data/redirects.ts | 9 ++++++++- 4 files changed, 22 insertions(+), 20 deletions(-) rename docs/cody/capabilities/{agentic-chat.mdx => agentic-context-fetching.mdx} (67%) diff --git a/docs/cody/capabilities/agentic-chat.mdx b/docs/cody/capabilities/agentic-context-fetching.mdx similarity index 67% rename from docs/cody/capabilities/agentic-chat.mdx rename to docs/cody/capabilities/agentic-context-fetching.mdx index 3b8c644e0..64cfa4cdb 100644 --- a/docs/cody/capabilities/agentic-chat.mdx +++ b/docs/cody/capabilities/agentic-context-fetching.mdx @@ -1,6 +1,6 @@ -# Agentic context fetching +# Agentic Context Fetching -

Learn about agentic context fetching, a mini-agent that uses search and tools to retrieve context.

+

Learn about agentic context fetching, a mini-agent that uses search and tools to retrieve context.

Cody's agentic context fetching experience can evaluate context and fetch any additional context (MCP, OpenCtx, terminal, etc.) by providing enhanced, context-aware chat capabilities. It extends Cody's functionality by proactively understanding your coding environment and gathering relevant information based on your requests before responding. These features help you get noticeably higher-quality responses. @@ -32,11 +32,11 @@ It has access to a suite of tools for retrieving relevant context. These tools i It integrates seamlessly with external services, such as web content retrieval and issue tracking systems, using OpenCtx providers. To learn more, [read the OpenCtx docs](/cody/capabilities/openctx). -Terminal access is not supported on the Web. It currently only works with VS Code, JetBrains, and Visual Studio editor extensions. +Terminal access is not supported on the Web. It currently only works with VS Code, JetBrains, and Visual Studio editor extensions. ## Terminal access -Agentic context fetching can use the CLI Tool to request the execution of shell commands to gather context from your terminal. Its ability to execute terminal commands enhances its context-gathering capabilities. However, it’s essential to understand that any information accessible via your terminal could potentially be shared with the LLM. It's recommended not to request information that you don't want to share. Here's what you should consider: +Agentic context fetching can use the CLI Tool to request the execution of shell commands to gather context from your terminal. Its ability to execute terminal commands enhances its context-gathering capabilities. However, it's essential to understand that any information accessible via your terminal could potentially be shared with the LLM. It's recommended not to request information that you don't want to share. Here's what you should consider: - **Requires user consent**: Agentic context fetching will pause and ask for permission each time before executing any shell command. - **Trusted workspaces only**: Commands can only be executed within trusted workspaces with a valid shell @@ -54,14 +54,8 @@ Agentic context fetching can be helpful to assist you with a wide range of tasks ## Enable agentic context fetching -### Getting agentic context fetching access for Pro users +Agentic context fetching is enabled by default for all Cody users. It uses LLM reflection and basic tool use steps to gather and refine context before sending it in the final model query. The review step in agentic context fetching experience defaults to Gemini 2.5 Flash and falls back to Claude Haiku or GPT 4.1 mini if Flash is unavailable. -Pro users can find the agentic context fetching option in the LLM selector drop-down. +If a user doesn't have any of these models, the agentic context fetching experience will use the model selected from the model dropdown for the final output. -![agentic context fetching interface](https://storage.googleapis.com/sourcegraph-assets/Docs/agenti%20-chat.jpg) - -### Getting agentic context fetching access for Enterprise customers - -Agentic context fetching uses smaller models from the Gemini, Claude, and GPT families for reflection steps and whichever model you choose from the model selector for the final response. This provide a good balance between quality and latency. If none of the smaller models are available on your instance, we will fall back to the model chosen in the model selector for reflection. We use the latest versions of these models, and can fall back to older versions when necessary. The default models may be changed to optimize for quality and/or latency. - -Terminal access is disabled by default. To enable it, set the `agentic-chat-cli-tool-experimental` feature flag [terminal access](#terminal-commands) +Terminal access for Enterprise users is disabled by default. To enable it, set the `agentic-chat-cli-tool-experimental` feature flag [terminal access](#terminal-commands). diff --git a/docs/cody/clients/feature-reference.mdx b/docs/cody/clients/feature-reference.mdx index 8dc9d61fa..01ac4b628 100644 --- a/docs/cody/clients/feature-reference.mdx +++ b/docs/cody/clients/feature-reference.mdx @@ -15,6 +15,7 @@ | @-file | ✅ | ✅ | ✅ | ✅ | ❌ | | @-symbol | ✅ | ❌ | ✅ | ✅ | ❌ | | LLM Selection | ✅ | ✅ | ✅ | ✅ | ❌ | +| Agentic Context Fetching | ✅ | ✅ | ✅ | ✅ | ✅ | | **Context Selection** | | | | | | | Single-repo context | ✅ | ✅ | ✅ | ✅ | ❌ | | Multi-repo context | ❌ | ❌ | ❌ | ✅ (public code only) | ❌ | @@ -26,11 +27,11 @@ ## Code Autocomplete and Auto-edit -| **Feature** | **VS Code** | **JetBrains** | **Visual Studio** | -| ------------------------------------------------------ | ----------- | ------------- | ----------------- | -| Single and multi-line autocompletion | ✅ | ✅ | ✅ | -| Cycle through multiple completion suggestions | ✅ | ✅ | ✅ | -| Accept suggestions word-by-word | ✅ | ❌ | ❌ | +| **Feature** | **VS Code** | **JetBrains** | **Visual Studio** | +| ----------------------------------------------------- | ----------- | ------------- | ----------------- | +| Single and multi-line autocompletion | ✅ | ✅ | ✅ | +| Cycle through multiple completion suggestions | ✅ | ✅ | ✅ | +| Accept suggestions word-by-word | ✅ | ❌ | ❌ | | Auto-edit suggestions via cursor movements and typing | ✅ | ✅ | ❌ | Few exceptions that apply to Cody Pro and Cody Enterprise users: diff --git a/src/data/navigation.ts b/src/data/navigation.ts index 0fd2f5bac..c6e2b0b8e 100644 --- a/src/data/navigation.ts +++ b/src/data/navigation.ts @@ -55,7 +55,7 @@ export const navigation: NavigationItem[] = [ subsections: [ { title: "Chat", href: "/cody/capabilities/chat", }, { title: "Query Types", href: "/cody/capabilities/query-types", }, - { title: "Agentic chat", href: "/cody/capabilities/agentic-chat", }, + { title: "Agentic Context Fetching", href: "/cody/capabilities/agentic-context-fetching", }, { title: "Autocomplete", href: "/cody/capabilities/autocomplete", }, { title: "Auto-edit", href: "/cody/capabilities/auto-edit", }, { title: "Prompts", href: "/cody/capabilities/prompts", }, diff --git a/src/data/redirects.ts b/src/data/redirects.ts index dd173546b..20464e746 100644 --- a/src/data/redirects.ts +++ b/src/data/redirects.ts @@ -6795,7 +6795,14 @@ const redirectsData = [ source: "/analytics/self-hosted", destination: "/analytics/air-gapped", permanent: true - } + }, + + //Agentic chat redirect + { + source: "/cody/capabilities/agentic-chat", + destination: "/cody/capabilities/agentic-context-fetching", + permanent: true + }, ]; From 59356b116879072e86281af2d4be42bd483b8530 Mon Sep 17 00:00:00 2001 From: Maedah Batool Date: Mon, 26 May 2025 20:44:22 -0700 Subject: [PATCH 04/11] Add docs for MCP support --- .../capabilities/agentic-context-fetching.mdx | 58 +++++++++++++++++++ 1 file changed, 58 insertions(+) diff --git a/docs/cody/capabilities/agentic-context-fetching.mdx b/docs/cody/capabilities/agentic-context-fetching.mdx index 64cfa4cdb..012ce030e 100644 --- a/docs/cody/capabilities/agentic-context-fetching.mdx +++ b/docs/cody/capabilities/agentic-context-fetching.mdx @@ -59,3 +59,61 @@ Agentic context fetching is enabled by default for all Cody users. It uses LLM r If a user doesn't have any of these models, the agentic context fetching experience will use the model selected from the model dropdown for the final output. Terminal access for Enterprise users is disabled by default. To enable it, set the `agentic-chat-cli-tool-experimental` feature flag [terminal access](#terminal-commands). + +## MCP support + +MCP (Model Context Protocol) tools are supported by agentic context fetching to connect to the external context. Users can configure multiple local MCP servers via their Cody's extension settings. Based on your query, agentic context will determine which MCP tools to invoke and what parameters to provide to the tools. Agentic context fetching experience will execute these tools and inject the context from the tool calls into Cody's context window. + +### Setting up MCP tools + +MCP is disabled by default. To enable it, add the `agentic-chat-mcp-enabled` feature flag to your Enterprise Sourcegraph instance. + +Once MCP is enabled, there are two ways to configure an MCP server: + +#### 1. From the editor extension UI + +To configure an MCP server from the editor extension UI, click your profile icon in the Cody extension and select the **MCP Settings** option. + +![mcp-via-ui](https://storage.googleapis.com/sourcegraph-assets/Docs/mcp-with-ui-2025.png) + +Create a new server configuration and input the necessary arguments provided by the server. You can disable individual tools in the UI by clicking on them. + +#### 2. From the editor extension settings + +- Edit your MCP configuration settings using the `cody.mcpServers` property in your extension settings. + - `settings.json` (VSCode) + - `cody_settings.json` (JetBrains) +- Use the following format when adding a new server + +In JetBrains, you need the absolute path to the command. Use the `which` command in the terminal to find it. For example, `which npx`. + +```json +"cody.mcpServers": { + "": { + "command": "...", + "args": [...], + "env": { + ... + }, + "disabledTools": [...] + } + } +``` + +When you submit a query, Cody will determine whether to use any of the server's tools. + +![mcp-via-settings](https://storage.googleapis.com/sourcegraph-assets/Docs/mcp-via-settings.png) + +### MCP Best Practices + +Since MCP is an open protocol and servers can be created by anyone, your mileage may vary. Tool names, descriptions, and the underlying models you have available may all affect the performance of MCP retrieval. To get the best results using your MCP servers, it's recommended the following: + +- Read through the tools of the servers you install and their parameters. Be as explicit as possible in your queries to provide all of the information the tool needs +- If you don't get the desired results on the first try, iterate on your query. It may take a couple of tries to get the results you're expecting +- Choose high-quality servers. Prefer servers from official sources over third parties. Read through the code to understand how the servers you are installing work + +### Limitations + +- Cody currently supports only local MCP servers +- MCP offers three main capabilities: Tools, Resources, and Prompts. Currently, Cody supports only Tools +- Since most MCP servers require users to create their access tokens, the servers will have the same access as whatever token you provide. For example, if a user has write permissions to Jira and creates an access token with write permissions, they can use Cody to write back into Jira via MCP. It is impossible to limit write access across all MCP tools broadly. You should use tools with this consideration in mind From a4da3cb4e046cb25de63551826c2421a7cbc8801 Mon Sep 17 00:00:00 2001 From: Maedah Batool Date: Tue, 27 May 2025 22:31:31 -0700 Subject: [PATCH 05/11] Update docs/cody/capabilities/agentic-context-fetching.mdx Co-authored-by: Alex Romano --- docs/cody/capabilities/agentic-context-fetching.mdx | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/cody/capabilities/agentic-context-fetching.mdx b/docs/cody/capabilities/agentic-context-fetching.mdx index 012ce030e..f71609b85 100644 --- a/docs/cody/capabilities/agentic-context-fetching.mdx +++ b/docs/cody/capabilities/agentic-context-fetching.mdx @@ -62,7 +62,7 @@ Terminal access for Enterprise users is disabled by default. To enable it, set t ## MCP support -MCP (Model Context Protocol) tools are supported by agentic context fetching to connect to the external context. Users can configure multiple local MCP servers via their Cody's extension settings. Based on your query, agentic context will determine which MCP tools to invoke and what parameters to provide to the tools. Agentic context fetching experience will execute these tools and inject the context from the tool calls into Cody's context window. +Cody supports [Model Context Protocol (MCP)](https://modelcontextprotocol.io/introduction) for connecting to external context. MCP servers are utilized via Cody’s agentic context fetching. Users can configure multiple local MCP servers via their Cody extension settings. Based on your query, agentic context will determine which MCP tools to invoke and what parameters to provide to the tools. Cody will execute these tools and inject the context from the tool calls into the context window. ### Setting up MCP tools From 475033ecec1827252c3a8714c80b1203c09d80ec Mon Sep 17 00:00:00 2001 From: Maedah Batool Date: Tue, 27 May 2025 22:31:40 -0700 Subject: [PATCH 06/11] Update docs/cody/capabilities/agentic-context-fetching.mdx Co-authored-by: Alex Romano --- docs/cody/capabilities/agentic-context-fetching.mdx | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/cody/capabilities/agentic-context-fetching.mdx b/docs/cody/capabilities/agentic-context-fetching.mdx index f71609b85..99a197840 100644 --- a/docs/cody/capabilities/agentic-context-fetching.mdx +++ b/docs/cody/capabilities/agentic-context-fetching.mdx @@ -66,7 +66,7 @@ Cody supports [Model Context Protocol (MCP)](https://modelcontextprotocol.io/int ### Setting up MCP tools -MCP is disabled by default. To enable it, add the `agentic-chat-mcp-enabled` feature flag to your Enterprise Sourcegraph instance. +MCP is disabled by default. To enable it, add the `agentic-context-mcp-enabled` feature flag to your Enterprise Sourcegraph instance. Once MCP is enabled, there are two ways to configure an MCP server: From 60b10eaa1219f1283c93ddef2c0fef51969ced09 Mon Sep 17 00:00:00 2001 From: Maedah Batool Date: Tue, 27 May 2025 22:31:50 -0700 Subject: [PATCH 07/11] Update docs/cody/capabilities/agentic-context-fetching.mdx Co-authored-by: Alex Romano --- docs/cody/capabilities/agentic-context-fetching.mdx | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/cody/capabilities/agentic-context-fetching.mdx b/docs/cody/capabilities/agentic-context-fetching.mdx index 99a197840..25094b372 100644 --- a/docs/cody/capabilities/agentic-context-fetching.mdx +++ b/docs/cody/capabilities/agentic-context-fetching.mdx @@ -106,7 +106,7 @@ When you submit a query, Cody will determine whether to use any of the server's ### MCP Best Practices -Since MCP is an open protocol and servers can be created by anyone, your mileage may vary. Tool names, descriptions, and the underlying models you have available may all affect the performance of MCP retrieval. To get the best results using your MCP servers, it's recommended the following: +Since MCP is an open protocol and servers can be created by anyone, your mileage may vary. Tool names, descriptions, and the underlying models you have available may all affect the performance of MCP retrieval. To get the best results using your MCP servers, we recommend the following: - Read through the tools of the servers you install and their parameters. Be as explicit as possible in your queries to provide all of the information the tool needs - If you don't get the desired results on the first try, iterate on your query. It may take a couple of tries to get the results you're expecting From f63decfa974ada9637467a0846a97defa627d248 Mon Sep 17 00:00:00 2001 From: Maedah Batool Date: Tue, 27 May 2025 22:31:59 -0700 Subject: [PATCH 08/11] Update docs/cody/capabilities/agentic-context-fetching.mdx Co-authored-by: Alex Romano --- docs/cody/capabilities/agentic-context-fetching.mdx | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/cody/capabilities/agentic-context-fetching.mdx b/docs/cody/capabilities/agentic-context-fetching.mdx index 25094b372..651bf7b87 100644 --- a/docs/cody/capabilities/agentic-context-fetching.mdx +++ b/docs/cody/capabilities/agentic-context-fetching.mdx @@ -116,4 +116,4 @@ Since MCP is an open protocol and servers can be created by anyone, your mileage - Cody currently supports only local MCP servers - MCP offers three main capabilities: Tools, Resources, and Prompts. Currently, Cody supports only Tools -- Since most MCP servers require users to create their access tokens, the servers will have the same access as whatever token you provide. For example, if a user has write permissions to Jira and creates an access token with write permissions, they can use Cody to write back into Jira via MCP. It is impossible to limit write access across all MCP tools broadly. You should use tools with this consideration in mind +- Since most MCP servers require users to create their access tokens, the servers will have the same access as whatever token you provide. For example, if a user has write permissions to Jira and creates an access token with write permissions, they can use Cody to write back into Jira via MCP. It is not possible today to limit write access across all MCP tools broadly. You should use tools with this consideration in mind From 31be4fb7340aa02262d77cf7499cca8bc7c2ff4d Mon Sep 17 00:00:00 2001 From: Maedah Batool Date: Tue, 27 May 2025 22:32:08 -0700 Subject: [PATCH 09/11] Update docs/cody/capabilities/agentic-context-fetching.mdx Co-authored-by: Alex Romano --- docs/cody/capabilities/agentic-context-fetching.mdx | 1 + 1 file changed, 1 insertion(+) diff --git a/docs/cody/capabilities/agentic-context-fetching.mdx b/docs/cody/capabilities/agentic-context-fetching.mdx index 651bf7b87..a19bb66f7 100644 --- a/docs/cody/capabilities/agentic-context-fetching.mdx +++ b/docs/cody/capabilities/agentic-context-fetching.mdx @@ -114,6 +114,7 @@ Since MCP is an open protocol and servers can be created by anyone, your mileage ### Limitations +- MCP is supported through Cody’s agentic context fetching capabilities. You must have these capabilities enabled in order to utilize MCP servers - Cody currently supports only local MCP servers - MCP offers three main capabilities: Tools, Resources, and Prompts. Currently, Cody supports only Tools - Since most MCP servers require users to create their access tokens, the servers will have the same access as whatever token you provide. For example, if a user has write permissions to Jira and creates an access token with write permissions, they can use Cody to write back into Jira via MCP. It is not possible today to limit write access across all MCP tools broadly. You should use tools with this consideration in mind From 80b12cb62790be965eee41d4eb385853a84eb733 Mon Sep 17 00:00:00 2001 From: Maedah Batool Date: Tue, 27 May 2025 22:32:14 -0700 Subject: [PATCH 10/11] Update docs/cody/capabilities/agentic-context-fetching.mdx Co-authored-by: Alex Romano --- docs/cody/capabilities/agentic-context-fetching.mdx | 2 ++ 1 file changed, 2 insertions(+) diff --git a/docs/cody/capabilities/agentic-context-fetching.mdx b/docs/cody/capabilities/agentic-context-fetching.mdx index a19bb66f7..f0c4d6b07 100644 --- a/docs/cody/capabilities/agentic-context-fetching.mdx +++ b/docs/cody/capabilities/agentic-context-fetching.mdx @@ -56,6 +56,8 @@ Agentic context fetching can be helpful to assist you with a wide range of tasks Agentic context fetching is enabled by default for all Cody users. It uses LLM reflection and basic tool use steps to gather and refine context before sending it in the final model query. The review step in agentic context fetching experience defaults to Gemini 2.5 Flash and falls back to Claude Haiku or GPT 4.1 mini if Flash is unavailable. +You can disable agentic context in your extension settings using `cody.agenticContext` + If a user doesn't have any of these models, the agentic context fetching experience will use the model selected from the model dropdown for the final output. Terminal access for Enterprise users is disabled by default. To enable it, set the `agentic-chat-cli-tool-experimental` feature flag [terminal access](#terminal-commands). From 18be9e82ee4b0ae01ce98cd7654499943ee6bd87 Mon Sep 17 00:00:00 2001 From: Maedah Batool Date: Tue, 27 May 2025 22:33:20 -0700 Subject: [PATCH 11/11] Add small tweaks --- docs/cody/capabilities/agentic-context-fetching.mdx | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/docs/cody/capabilities/agentic-context-fetching.mdx b/docs/cody/capabilities/agentic-context-fetching.mdx index f0c4d6b07..73fac727a 100644 --- a/docs/cody/capabilities/agentic-context-fetching.mdx +++ b/docs/cody/capabilities/agentic-context-fetching.mdx @@ -56,9 +56,7 @@ Agentic context fetching can be helpful to assist you with a wide range of tasks Agentic context fetching is enabled by default for all Cody users. It uses LLM reflection and basic tool use steps to gather and refine context before sending it in the final model query. The review step in agentic context fetching experience defaults to Gemini 2.5 Flash and falls back to Claude Haiku or GPT 4.1 mini if Flash is unavailable. -You can disable agentic context in your extension settings using `cody.agenticContext` - -If a user doesn't have any of these models, the agentic context fetching experience will use the model selected from the model dropdown for the final output. +You can disable agentic context in your extension settings using `cody.agenticContext`. Terminal access for Enterprise users is disabled by default. To enable it, set the `agentic-chat-cli-tool-experimental` feature flag [terminal access](#terminal-commands).