Skip to main content
cubic’s MCP server connects your coding agent to cubic’s data — review issues on your PRs, AI-generated codebase documentation, codebase scan results, and team review patterns. Instead of switching to GitHub or the cubic dashboard, your agent can query this information directly while you code.
The MCP server requires a cubic account with access to at least one repository. You must be a member of the repository’s installation to access its data.

​
What you can do

Once connected, your agent can:
  • Fetch PR review issues — Get all open cubic review comments for a pull request, grouped by file with severity and categories
  • Query codebase documentation — Access AI Wiki pages containing architecture overviews, feature explanations, and code references
  • Review codebase scan results — List scans, inspect issues, and pull detailed findings with code context
  • Inspect review learnings — Understand team-specific review patterns learned from feedback

​
Quick setup

1

Generate an API key

Go to Settings → Integrations → MCP Configuration and click Generate API Key. Copy the key immediately — it’s only shown once.
API keys are personal to each user. Each team member generates their own key, which grants access only to repositories they’re authorized to view.
MCP Server configuration showing API key and client setup
2

Configure your AI client

Add the cubic MCP server to your coding agent. See Coding agent integration for setup instructions, or use the generic configuration below.
3

Start using MCP tools

Ask your agent to check cubic comments on a PR, browse wiki documentation, or inspect scan findings for your repositories.

​
Client configuration

For agent-specific MCP setup instructions, see Coding agent integration. Each agent tab includes manual MCP configuration.

​
Generic configuration

For MCP clients not listed on the agent setup page, use these connection details:
SettingValue
URLhttps://www.cubic.dev/api/mcp
TransportHTTP (Streamable)
AuthenticationBearer token in Authorization header

​
Available tools

​
PR review tools

Get AI code review issues posted on your pull requests. Use this to see what cubic flagged, understand issue severity, and check which comments still need attention.
ToolDescriptionParameters
get_pr_issuesGet all open review issues for a pull requestrepo, pullNumber, owner (optional)
Recommended workflow:
  1. Call get_pr_issues with a repo name and PR number
  2. Review the issues grouped by file, with severity and confidence scores
  3. Address the issues in your editor
The owner parameter is optional — omit it if you only have the repo name (e.g., from a PR URL like github.com/org/repo/pull/123).
Ask your agent:
“Show me the cubic review issues on PR #42 in the backend repo”
The agent calls get_pr_issues and returns a report like:
# AI Code Review: acme/backend PR #42

| Property    | Value |
|-------------|-------|
| Open Issues | 3     |

## src/api/auth.ts (2 issues)

### 1. High — L15-20
- Severity: 8/10 | Confidence: 9/10
- Categories: security

Missing rate limiting on the login endpoint...

### 2. Medium — L45
- Severity: 5/10 | Confidence: 7/10
- Categories: stability

Error not propagated to caller...

​
Wiki tools

AI Wikis are AI-generated codebase documentation containing architecture overviews, feature explanations, code references, and diagrams. Use these tools to explore and understand codebases without leaving your editor.
Enable auto-refresh on your wiki to keep documentation up to date as your codebase evolves.
ToolDescriptionParameters
list_wikisList all wikis you have access tosearch (optional)
list_wiki_pagesList all available wiki pages for a repositoryowner, repo
get_wiki_pageRetrieve the content of a specific wiki pageowner, repo, pageRef
Recommended workflow:
  1. Call list_wikis to discover repositories with generated wikis
  2. Call list_wiki_pages to see available documentation for a repository
  3. Use get_wiki_page with a page reference to read specific content
Ask your agent:
“List the wiki pages for the acme/backend repository”
The agent calls list_wiki_pages and returns:
Available wiki pages for acme/backend:
1. architecture-overview
2. authentication-system
3. database-schema
4. api-endpoints
5. deployment-guide
Then ask:
“Show me the authentication system documentation”
The agent calls get_wiki_page with pageRef: "authentication-system" and returns the full documentation.

​
Codebase scan tools

Codebase scans identify security vulnerabilities, data integrity issues, business logic bugs, and stability concerns across your repositories.
ToolDescriptionParameters
list_scansList repository scan summariessearch?, limit?, offset?
get_scanList aggregated issues with filters/paginationowner, repo, category?, triageStatus?, minSeverity?, filePath?, limit?, offset?
get_issueGet full issue details including code contextissueId
Recommended workflow:
  1. Call list_scans to discover repositories with scan results and open issue counts
  2. Call get_scan with owner/repo to list issues — optionally filter by category, severity, or file path
  3. Call get_issue for complete details on a specific issue, including code context and severity scores
get_scan supports these filters:
  • category — security, data_integrity, business_logic, or stability
  • triageStatus — open, in_review, resolved, false_positive, wont_fix, intended_behavior, or accepted_risk
  • minSeverity — Inclusive, 0 to 10
  • filePath — Case-sensitive prefix match (e.g., src/app/api)
Default page size is 20, server cap is 100. Use offset for pagination.

​
Review learning tools

AI review learnings are team-specific code review patterns learned from senior reviewer feedback, user corrections, and codebase scans. Use these tools to understand what review preferences apply to a repository.
ToolDescriptionParameters
list_learningsList all AI review learnings for a repositoryowner, repo
get_learningGet full details of a specific learninglearningId
Recommended workflow:
  1. Call list_learnings to discover patterns and preferences for a repository
  2. Call get_learning with a learning ID for full details and original feedback context

​
Troubleshooting

  • Your API key is correct and hasn’t been revoked
  • You’re using the correct endpoint: https://www.cubic.dev/api/mcp
  • Your firewall or network allows outbound HTTPS connections
  • Check that your API key is prefixed with cbk_
  • Ensure the Authorization header uses Bearer format
  • Regenerate your API key if issues persist
  • Confirm the repository has an AI Wiki generated
  • Verify you have access to the repository’s installation
  • Check that the owner and repo parameters match the GitHub repository exactly
  • Restart your AI client after updating the configuration
  • Check the config file path is correct for your operating system
  • Validate your JSON syntax (no trailing commas, proper quoting)