github.com/Checkmk/checkmk (link)

commit signals (181)

source signals (17)

• .github/copilot-instructions.mdvia github-copilot

  # Copilot instructions for check_mk
  
  Project conventions are documented in `CLAUDE.md` at the repository root —
  apply those when suggesting changes.
  
  ## AI tooling lives elsewhere
  
  Reusable AI workflows for this project (code review, werk authoring,
  Gerrit/Jenkins/Jira/crash-report helpers, backports, etc.) are maintained
  as Claude Code plugins in a separate, internal repository:
  
      https://github.com/Checkmk/checkmk-claude-marketplace
  
  Copilot cannot execute these plugins, but their `SKILL.md` files de

• .github/skillsdirectoryvia github-copilot
• .github/agentsdirectoryvia github-copilot
• .mcp.jsonvia unspecified

  {
    "mcpServers": {
      "cmk-frontend-vue--ui-component-library": {
        "command": "bazel",
        "args": ["run", "//packages/cmk-frontend-vue:ucl-mcp-server"]
      }
    }
  }
  

• .claude/skills/cmk-up/SKILL.mdvia unspecified

  ---
  name: cmk-up
  description: Writes executable cmk-up playbooks (YAML that builds Checkmk sites, hosts, services, rules, BI on local OMD) that reproduce a scenario - from a Jira ticket (SUP-/CMK-/...), a scenario described in prose, or a Checkmk crash report (tarball, crash dir, crash.info, crash.checkmk.com id). Use when asked to "reproduce", "write a playbook", "cmk-up", "repro for SUP-…", "playbook from crash".
  ---
  
  # cmk-up playbooks
  
  cmk-up (package `cmk-werk-zeug`) reads a YAML playbook, installs v

• AGENTS.mdvia agents-md

  # AI Agent Instructions
  
  ## Build & Test
  
  **NEVER run `pytest`, `ruff`, or `mypy` directly.** Always use `bazel` commands
  instead. Direct invocations bypass the build system's environment and dependency
  setup and produce incorrect results.
  
  Bazel is the primary build system for unit tests, linting, formatting, and type
  checking. System tests use run_tests.sh:
  
  ```
  tests/run_tests.sh test-system-singlesite
  tests/run_tests.sh test-system-multisite
  tests/run_tests.sh test-system-gui
  ```
  
  Always format, lint, a

• packages/cmk-agent-receiver/CLAUDE.mdvia claude-code

  # cmk-agent-receiver
  
  See [README.md](README.md) for architecture, API, configuration, and development workflows.
  
  ## Testing
  
  Component tests offer two app-startup modes — pick based on what you're testing:
  
  - **`TestClient`** (default fixture): in-process, fast, covers most endpoint logic
  - **`AgentReceiverRunner`**: spawns a real Gunicorn process; required for anything involving mTLS or the `ClientCertWorker` certificate extraction
  

• CLAUDE.mdvia claude-code

  AGENTS.md

• .ide/vscode/CLAUDE.mdvia claude-code

  # CMK VS Code Extension — AI Agent Instructions
  
  ## Quick Reference
  
  - **Language**: TypeScript, bundled via esbuild into a single CommonJS file
  - **Entry point**: `src/extension.ts` → compiled to `out/extension.js`
  - **UI**: Sidebar sections are webviews rendered by per-section modules in `src/sidebar/`
  - **Config-driven**: JSON files in `config/` define commands, extensions, settings, snippets, templates
  - **Config loading**: Workspace-first (`<workspace>/.ide/vscode/config/`), falls back to bundled V

• .claudedirectoryvia claude-code
• .claudedirectoryvia claude-code
• .claudedirectoryvia claude-code
• .claudedirectoryvia claude-code
• .claudedirectoryvia claude-code
• .claudedirectoryvia claude-code
• .claudedirectoryvia claude-code
• .claudedirectoryvia claude-code

ignorefile signals (2)