|
1 | 1 | --- |
2 | | -description: Monitors and updates agentic CLI tools (Claude Code, GitHub Copilot CLI, OpenAI Codex, GitHub MCP Server, Playwright MCP, Playwright Browser) for new versions |
| 2 | +description: Monitors and updates agentic CLI tools (Claude Code, GitHub Copilot CLI, OpenAI Codex, GitHub MCP Server, Playwright MCP, Playwright Browser, Sandbox Runtime) for new versions |
3 | 3 | on: |
4 | 4 | schedule: daily |
5 | 5 | workflow_dispatch: |
@@ -28,7 +28,7 @@ timeout-minutes: 45 |
28 | 28 |
|
29 | 29 | # CLI Version Checker |
30 | 30 |
|
31 | | -Monitor and update agentic CLI tools: Claude Code, GitHub Copilot CLI, OpenAI Codex, GitHub MCP Server, Playwright MCP, and Playwright Browser. |
| 31 | +Monitor and update agentic CLI tools: Claude Code, GitHub Copilot CLI, OpenAI Codex, GitHub MCP Server, Playwright MCP, Playwright Browser, and Sandbox Runtime. |
32 | 32 |
|
33 | 33 | **Repository**: ${{ github.repository }} | **Run**: ${{ github.run_id }} |
34 | 34 |
|
@@ -62,6 +62,9 @@ For each CLI/MCP server: |
62 | 62 | - **Playwright Browser**: `https://api.github.com/repos/microsoft/playwright/releases/latest` |
63 | 63 | - Release Notes: https://github.com/microsoft/playwright/releases |
64 | 64 | - Docker Image: `mcr.microsoft.com/playwright:v{VERSION}` |
| 65 | +- **Sandbox Runtime**: Use `npm view @anthropic-ai/sandbox-runtime version` |
| 66 | + - Repository: https://github.com/anthropic-experimental/sandbox-runtime |
| 67 | + - Package: https://www.npmjs.com/package/@anthropic-ai/sandbox-runtime |
65 | 68 |
|
66 | 69 | **Optimization**: Fetch all versions in parallel using multiple npm view or WebFetch calls in a single turn. |
67 | 70 |
|
@@ -105,11 +108,13 @@ For each CLI tool update: |
105 | 108 | - Copilot CLI: `npm install -g @github/copilot@<version>` |
106 | 109 | - Codex: `npm install -g @openai/codex@<version>` |
107 | 110 | - Playwright MCP: `npm install -g @playwright/mcp@<version>` |
| 111 | + - Sandbox Runtime: `npm install -g @anthropic-ai/sandbox-runtime@<version>` |
108 | 112 | 2. Invoke help to discover commands and flags (compare with cached output if available): |
109 | 113 | - Run `claude-code --help` |
110 | 114 | - Run `copilot --help` or `copilot help copilot` |
111 | 115 | - Run `codex --help` |
112 | 116 | - Run `npx @playwright/mcp@<version> --help` (if available) |
| 117 | + - Sandbox Runtime is a library, check NPM package metadata for API changes |
113 | 118 | 3. **Explore subcommand help** for each tool (especially Copilot CLI): |
114 | 119 | - Identify all available subcommands from main help output |
115 | 120 | - For each subcommand, run its help command (e.g., `copilot help config`, `copilot help environment`, `copilot config --help`) |
|
0 commit comments