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

# Developer Tools

> Integrate Kasava's product intelligence into your IDE and AI coding agents

Kasava's developer tools bring product intelligence directly into your coding workflow. Search code, query your product graph, load plan specs, and check customer signals — without leaving your editor.

## Choose Your Setup

<CardGroup cols={2}>
  <Card title="Claude Code Plugin" icon="bolt" href="/docs/developer/claude-code-plugin">
    One-click install with 6 auto-trigger skills, 8 slash commands, 4 subagents, and 8 MCP tools. The fastest way to get started in Claude Code.
  </Card>

  <Card title="Kasava CLI" icon="terminal" href="/docs/developer/cli">
    Bind a repo to a Kasava product and sync plans, graph, and symbols into a checked-in `.kasava/` artifact. Terminal-first companion.
  </Card>

  <Card title="MCP Server" icon="plug" href="/docs/developer/mcp-setup">
    Connect any MCP-compatible client (Cursor, Claude Desktop, Windsurf) to Kasava's 8 live-data tools.
  </Card>

  <Card title="Agent Skills" icon="brain" href="/docs/developer/agent-skills">
    Portable skills that work in Claude Code, Cursor, Copilot, Codex, Gemini CLI, and 30+ tools.
  </Card>

  <Card title="MCP Tool Reference" icon="book" href="/docs/developer/api-reference">
    Complete reference for all available MCP tools, actions, and parameters.
  </Card>
</CardGroup>

## What You Get

### 8 MCP Tools (live data)

The MCP server exposes 8 live-data tools (cut from 23 in May 2026 — cacheable data moved to the local `.kasava/` artifact). When connected, your AI coding agent can:

* **Search code** using Kasava's symbol graph — definitions, usages, dependencies, semantic search
* **Search across signals + plans + docs** with the unified search tool
* **Analyze commits** with semantic classification, hotspots, and changelog
* **Get competitor intel** — live scrapings, change detection, market trends
* **Traverse the product graph** for entity relationships and architecture
* **Export plan specs** in agent-consumable format
* **Record decisions** against the bound product (only write tool)

Plus 3 server-side AI agents (`ask_*`) for code intelligence, product insights, and architecture investigation, and 2 multi-step workflows (`run_*`) for bug analysis and deep research.

### 6 Agent Skills

Auto-trigger workflows that teach AI agents how to use Kasava effectively:

| Skill            | Purpose                                                          |
| ---------------- | ---------------------------------------------------------------- |
| `kasava-context` | Read the local `.kasava/` artifact (paths-scoped to bound repos) |
| `kasava-product` | Load architecture, features, tech stack                          |
| `kasava-code`    | Navigate codebases via the symbol graph                          |
| `kasava-bug`     | 3-way investigation: code + commits + signals                    |
| `kasava-plan`    | Pull plan specs and resolve them to current code                 |
| `kasava-signals` | Ground decisions in customer feedback                            |

Plus 8 user-invoked slash commands (`/kasava:install`, `:sync`, `:plan`, `:ask`, `:ask-ai`, `:staleness`, `:decision`, `:status`) and 4 subagents (`kasava-developer`, `kasava-bug-triager`, `kasava-spec-loader`, `kasava-impact-assessor`) — see the [Claude Code Plugin guide](/docs/developer/claude-code-plugin) for details.

### Prerequisites

* A Kasava account with at least one indexed repository
* An API key ([Settings > API Keys](https://app.kasava.dev/settings/api-keys))