> ## Documentation Index
> Fetch the complete documentation index at: https://docs.endorlabs.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Endor Labs Agent Kit

> Ready-to-use Endor Labs security agents for AI coding assistants.

The Endor Labs Agent Kit is a catalog of ready-to-use security agents that run inside your AI coding assistant. Each agent packages a specific Endor Labs workflow, such as triaging findings, remediating vulnerable dependencies, or diagnosing scan failures, with a built-in safety contract, approval gates, and evidence requirements. Install the kit as a plugin or extension in your host of choice, then ask the assistant to run a workflow in plain language.

The Agent Kit replaces the earlier Endor Labs Skills offering, which could only install and configure `endorctl`. The kit ships a full set of workflow agents across every supported host.

## How it is distributed

The Agent Kit lives in two open-source repositories:

* **[endorlabs/ai-plugins](https://github.com/endorlabs/ai-plugins)**: The distribution and marketplace repository. This is what you install from. It is already published and branded in the Claude Code and Cursor marketplaces.
* **[endorlabs/endor-labs-agent-kit](https://github.com/endorlabs/endor-labs-agent-kit)**: The source and builder repository. Use it to contribute or propose new agents, file issues, or grab runtime-neutral [portable agents](https://docs.endorlabs.com/secure-ai-coding/agent-kit/portable.md) for your own agent runtime.

> [!NOTE]
>
> **Upgrading from Endor Labs Skills**
>
> If you previously installed the Endor Labs plugin as `ai-plugins@endorlabs`, it keeps working and existing installs are unaffected. New users should install the `endor-labs-agent-kit@endorlabs` plugin instead. Do not enable both plugin IDs in the same Claude Code profile, because they expose the same agents and setup skill.

## Supported hosts

Install the kit in any of the following hosts.


&#x20; \- Host: Claude Code
&#x20; How\_to\_install: Plugin from the Claude Code plugin marketplace
&#x20; \- Host: OpenAI Codex
&#x20; How\_to\_install: Plugin from the Codex plugin marketplace
&#x20; \- Host: Gemini CLI
&#x20; How\_to\_install: Gemini extension
&#x20; \- Host: Antigravity CLI
&#x20; How\_to\_install: Antigravity plugin
&#x20; \- Host: Cursor
&#x20; How\_to\_install: Cursor plugin for the IDE
&#x20; \- Host: Cursor SDK
&#x20; How\_to\_install: Python SDK for automation, CI, and cloud runs
&#x20; \- Host: Portable
&#x20; How\_to\_install: Runtime-neutral bundles for your own agent runtime
&#x20;&#x20;

## Agent catalog

The kit ships 11 agents. Most are read-only. Two agents, AI SAST Remediation and SCA Remediation, can change state, and they keep every mutating action behind a separate approval gate. See the [safety model](https://docs.endorlabs.com/secure-ai-coding/agent-kit/safety-model.md) for details.

> [!TIP]
>
> In a hurry? Jump to [Quick start](#quick-start) and pick your host and agent to copy the exact install command and first-workflow prompt.

If you are new to the Agent Kit, start with the four most popular agents: **SCA Remediation**, **AI SAST Remediation**, **Configuration Automation**, and **Troubleshooting**.

The **License** column shows the Endor Labs license each agent requires. Agents marked **Any** work with any Endor Labs license because they operate on findings and package intelligence you already have. For what each license includes, see [Licenses](https://docs.endorlabs.com/introduction/licenses.md).


&#x20; \- Agent: SCA Remediation
&#x20; Workflow: \`sca-remediation\`
&#x20; What\_it\_does: Plan and apply approval-gated dependency fixes using Endor Labs SCA findings, upgrade-risk evidence, and local validation.
&#x20; Safety: mutating, approval-gated
&#x20; License: Endor Open Source Pro
&#x20; \- Agent: AI SAST Remediation
&#x20; Workflow: \`ai-sast-remediation\`
&#x20; What\_it\_does: Triage and remediate Endor Labs AI SAST findings using exploit-reproduction evidence, data-flow context, and approval-gated fixes.
&#x20; Safety: mutating, approval-gated
&#x20; License: Endor Code Pro
&#x20; \- Agent: Configuration Automation
&#x20; Workflow: \`configuration-automation\`
&#x20; What\_it\_does: Find GitHub-to-Endor Labs onboarding and monitored-branch coverage gaps, and prescribe the setup to close them, without changing anything.
&#x20; Safety: read-only
&#x20; License: Any (GitHub only)
&#x20; \- Agent: Troubleshooting
&#x20; Workflow: \`troubleshooting\`
&#x20; What\_it\_does: Diagnose Endor Labs setup and workflow problems, such as scan failures and missing integrations, from focused read-only evidence.
&#x20; Safety: read-only
&#x20; License: Any
&#x20; \- Agent: CI/CD And Supply Chain Posture
&#x20; Workflow: \`cicd-posture\`
&#x20; What\_it\_does: Score CI/CD and supply chain posture from read-only Endor Labs findings and repository configuration evidence.
&#x20; Safety: read-only
&#x20; License: Endor Open Source Pro
&#x20; \- Agent: Dependency Reviewer
&#x20; Workflow: \`dependency-reviewer\`
&#x20; What\_it\_does: Review an exact package version, a package's risk profile, or a repository's declared dependencies through one bounded workflow.
&#x20; Safety: read-only
&#x20; License: Any
&#x20; \- Agent: OSS Upgrade Investigator
&#x20; Workflow: \`oss-upgrade-investigator\`
&#x20; What\_it\_does: Compare Endor Labs upgrade candidates, risk, breaking changes, and code impact to identify the safest upgrade path.
&#x20; Safety: read-only
&#x20; License: Endor Open Source Pro
&#x20; \- Agent: Vulnerability Explainer
&#x20; Workflow: \`vulnerability-explainer\`
&#x20; What\_it\_does: Explain a vulnerability's severity, exploitability, affected versions, and recommended remediation.
&#x20; Safety: read-only
&#x20; License: Any
&#x20; \- Agent: Findings Browser
&#x20; Workflow: \`findings-browser\`
&#x20; What\_it\_does: Browse and filter existing Endor Labs findings with clear scope, pagination limits, and evidence gaps.
&#x20; Safety: read-only
&#x20; License: Any
&#x20; \- Agent: Malware Responder
&#x20; Workflow: \`malware-responder\`
&#x20; What\_it\_does: Correlate current supply chain malware intelligence with your Endor Labs inventory to assess tenant exposure.
&#x20; Safety: read-only
&#x20; License: Any
&#x20; \- Agent: Remediation Planning
&#x20; Workflow: \`remediation-planning\`
&#x20; What\_it\_does: Compare read-only remediation options and recommend the safest evidence-backed next step without opening pull requests.
&#x20; Safety: read-only
&#x20; License: Endor Open Source Pro
&#x20;&#x20;

Each host exposes these workflows with its own invocation syntax. See the install page for your host for exact commands and example prompts.

## Quick start

Choose your host and a starting agent to generate the matching install command, setup prompt, and first-workflow prompt. The four most popular agents appear as buttons. Select **Other agents** to pick any of the remaining agents from the dropdown.

Replace placeholders such as \<org> and \<namespace> with your own values before running.

Using Cursor or running agents programmatically? See [Cursor](https://docs.endorlabs.com/secure-ai-coding/agent-kit/cursor.md) and the [Cursor SDK](https://docs.endorlabs.com/secure-ai-coding/agent-kit/cursor-sdk.md).

## Run setup first

After you install the kit in any host, run the setup skill before anything else.

```text theme={null}
Use the endor-agent-kit-setup skill to check Endor Labs Agent Kit readiness.
```

Setup is readiness guidance only. It checks for `endorctl`, authentication, namespace selection, `gh`, and toolchain prerequisites, and reports what is missing. It never runs scans, runs `endorctl host-check`, edits shell profiles, installs runtimes, or writes credentials.

## Choose your host to get started

> [card] [Claude Code](https://docs.endorlabs.com/secure-ai-coding/agent-kit/claude-code.md)
>
> Install the Endor Labs Agent Kit plugin in Claude Code.

> [card] [Cursor](https://docs.endorlabs.com/secure-ai-coding/agent-kit/cursor.md)
>
> Install the Endor Labs Agent Kit plugin in Cursor.

> [card] [OpenAI Codex](https://docs.endorlabs.com/secure-ai-coding/agent-kit/codex.md)
>
> Install the Endor Labs Agent Kit plugin in OpenAI Codex.

> [card] [Gemini CLI](https://docs.endorlabs.com/secure-ai-coding/agent-kit/gemini-cli.md)
>
> Install the Endor Labs Agent Kit extension in Gemini CLI.

> [card] [Antigravity CLI](https://docs.endorlabs.com/secure-ai-coding/agent-kit/antigravity-cli.md)
>
> Install the Endor Labs Agent Kit plugin in Antigravity CLI.

> [card] [Cursor SDK](https://docs.endorlabs.com/secure-ai-coding/agent-kit/cursor-sdk.md)
>
> Run Agent Kit workflows programmatically with the Cursor Python SDK.

> [card] [Portable agents](https://docs.endorlabs.com/secure-ai-coding/agent-kit/portable.md)
>
> Run runtime-neutral agents in your own agent runtime.

> [card] [Safety model](https://docs.endorlabs.com/secure-ai-coding/agent-kit/safety-model.md)
>
> Understand the safety contract and output guarantees.

To scan your code for vulnerabilities, secrets, and SAST issues directly in your IDE, see the [Endor Labs MCP server](https://docs.endorlabs.com/setup-deployment/mcp.md).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.