greenrun-cli 0.1.2 → 0.1.4

package/dist/cli.js

@@ -15,6 +15,7 @@ function printHelp() {
 
 Usage:
   greenrun init             Interactive setup wizard
+  greenrun update           Update command templates to latest version
   greenrun serve            Start MCP server
   greenrun --version, -v    Print version
   greenrun --help, -h       Print this help
@@ -42,6 +43,11 @@ async function main() {
         await runInit(args.slice(1));
         return;
     }
+    if (command === 'update') {
+        const { runUpdate } = await import('./commands/init.js');
+        runUpdate();
+        return;
+    }
     if (command === 'serve') {
         const { startServer } = await import('./server.js');
         await startServer();

package/dist/commands/init.d.ts

@@ -1 +1,2 @@
+export declare function runUpdate(): void;
 export declare function runInit(args: string[]): Promise<void>;

package/dist/commands/init.js

@@ -153,6 +153,12 @@ function installCommands() {
         console.log(`  Installed /${cmd.replace('.md', '')}`);
     }
 }
+export function runUpdate() {
+    console.log('\nGreenrun - Updating templates\n');
+    installCommands();
+    installClaudeMd();
+    console.log('\nDone! Templates updated to latest version.\n');
+}
 export async function runInit(args) {
     const opts = parseFlags(args);
     const interactive = !opts.token;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "greenrun-cli",
-  "version": "0.1.2",
+  "version": "0.1.4",
   "description": "CLI and MCP server for Greenrun - browser test management for Claude Code",
   "type": "module",
   "main": "dist/server.js",

package/templates/commands/greenrun-sweep.md

@@ -40,7 +40,7 @@ Ask the user if they want to run the affected tests. If yes, execute them **in p
 
 Use the project's `concurrency` setting (default: 5) to determine batch size. Split affected tests into batches and launch each batch simultaneously using the **Task tool** with `run_in_background: true`.
 
-For each test in a batch, launch a background agent with this prompt:
+For each test in a batch, launch a background agent with `max_turns: 30` and `model: "sonnet"`. Use this prompt:
 
 ```
 You are executing a single Greenrun browser test. You have access to browser automation tools and Greenrun MCP tools.
@@ -50,14 +50,19 @@ You are executing a single Greenrun browser test. You have access to browser aut
 Step 1: Call `get_test` with test_id "{test_id}" to get full instructions.
 Step 2: Call `start_run` with test_id "{test_id}" to begin - save the returned `run_id`.
 Step 3: Execute the test instructions using browser automation:
-   - Create a new browser tab for this test
+   - Call `tabs_context_mcp` then create a new browser tab for this test
    - Follow each instruction step exactly as written
    - The instructions will tell you where to navigate and what to do
-   - Observe results and take screenshots as needed for verification
+   - Only take a screenshot when you need to verify a visual assertion — not for every navigation or click
+   - When reading page content, prefer `find` or `read_page` with `filter: "interactive"` over full DOM reads
+   - NEVER trigger JavaScript alerts, confirms, or prompts — they block the browser extension entirely. Before clicking delete buttons or other destructive actions, use `javascript_tool` to override: `window.alert = () => {}; window.confirm = () => true; window.prompt = () => null;`
+   - If browser tools stop responding (no result or timeout), assume a dialog is blocking — report the error and stop. Do not keep retrying.
+   - If you get stuck or a step fails, record the failure and move on — do not retry more than once
 Step 4: Call `complete_run` with:
    - run_id: the run ID from step 2
    - status: "passed" if all checks succeeded, "failed" if any check failed, "error" if execution was blocked
-   - result: a brief summary of what happened
+   - result: a brief summary of what happened (include the failure reason if failed/error)
+Step 5: Close the browser tab you created to clean up.
 
 Return a single line summary: {test_name} | {status} | {result_summary}
 ```

package/templates/commands/greenrun.md

@@ -33,6 +33,8 @@ For each batch, launch all tests simultaneously using the **Task tool** with `ru
 For each test in the current batch, call the Task tool with:
 - subagent_type: "general-purpose"
 - run_in_background: true
+- max_turns: 30
+- model: "sonnet"
 - prompt: (see below)
 ```
 
@@ -46,14 +48,19 @@ You are executing a single Greenrun browser test. You have access to browser aut
 Step 1: Call `get_test` with test_id "{test_id}" to get full instructions.
 Step 2: Call `start_run` with test_id "{test_id}" to begin - save the returned `run_id`.
 Step 3: Execute the test instructions using browser automation:
-   - Create a new browser tab for this test
+   - Call `tabs_context_mcp` then create a new browser tab for this test
    - Follow each instruction step exactly as written
    - The instructions will tell you where to navigate and what to do
-   - Observe results and take screenshots as needed for verification
+   - Only take a screenshot when you need to verify a visual assertion — not for every navigation or click
+   - When reading page content, prefer `find` or `read_page` with `filter: "interactive"` over full DOM reads
+   - NEVER trigger JavaScript alerts, confirms, or prompts — they block the browser extension entirely. Before clicking delete buttons or other destructive actions, use `javascript_tool` to override: `window.alert = () => {}; window.confirm = () => true; window.prompt = () => null;`
+   - If browser tools stop responding (no result or timeout), assume a dialog is blocking — report the error and stop. Do not keep retrying.
+   - If you get stuck or a step fails, record the failure and move on — do not retry more than once
 Step 4: Call `complete_run` with:
    - run_id: the run ID from step 2
    - status: "passed" if all checks succeeded, "failed" if any check failed, "error" if execution was blocked
-   - result: a brief summary of what happened
+   - result: a brief summary of what happened (include the failure reason if failed/error)
+Step 5: Close the browser tab you created to clean up.
 
 Return a single line summary: {test_name} | {status} | {result_summary}
 ```