agentic-loop 3.2.0 → 3.2.2

package/.claude/commands/idea.md

@@ -306,7 +306,7 @@ For frontend stories, acceptance criteria MUST include:
 - "Required elements render" (specify which: header, form, button, etc.)
 - "Works on mobile viewport (375px)"
 
-These get verified by Playwright, not just code review.
+These get verified by Playwright tests and MCP browser tools.
 
 ### Test Steps - CRITICAL
 **Test steps MUST be executable shell commands.** Ralph runs them with bash.
@@ -352,7 +352,7 @@ The Playwright test file can check:
 ]
 ```
 
-**If a step can't be automated**, leave it out of testSteps and put it in acceptanceCriteria instead. Ralph will verify acceptanceCriteria via code review, not by running commands.
+**If a step can't be automated**, leave it out of testSteps and put it in acceptanceCriteria instead. Claude will verify acceptanceCriteria visually using MCP browser tools.
 
 ### Architecture Guidelines
 - **Domain-driven directories** - Group by feature (`src/contact/`) not type (`src/components/`)

package/.claude/commands/prd.md

@@ -283,4 +283,4 @@ The Playwright test file can check:
 ]
 ```
 
-**If a step can't be automated**, leave it out of testSteps and put it in acceptanceCriteria instead. Ralph will verify acceptanceCriteria via code review, not by running commands.
+**If a step can't be automated**, leave it out of testSteps and put it in acceptanceCriteria instead. Claude will verify acceptanceCriteria visually using MCP browser tools.

package/README.md

@@ -45,8 +45,7 @@ claude --dangerously-skip-permissions
 
 **Terminal 2 - Ralph Loop:**
 ```bash
-npx agentic-loop run       # Execute PRDs autonomously
-npx agentic-loop run --fast  # Skip code review (~2x faster)
+npx agentic-loop run         # Execute PRDs autonomously
 ```
 
 > **Tip:** Use two terminals. Plan with Claude in one, run Ralph in the other.
@@ -61,8 +60,8 @@ npx agentic-loop run --fast  # Skip code review (~2x faster)
 ├─────────────────────────────────────────────────────────────┤
 │  1. Read prd.json → find next story where passes=false      │
 │  2. Build prompt (story + context + failures + signs)       │
-│  3. Spawn Claude with prompt                                │
-│  4. Run verification (lint, tests, browser, code review)    │
+│  3. Spawn Claude with prompt + MCP browser tools            │
+│  4. Run verification (lint, tests, testSteps)               │
 │  5. Pass? → commit, next story                              │
 │     Fail? → save error, retry with failure context          │
 │  6. Repeat until all stories pass                           │

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "agentic-loop",
-  "version": "3.2.0",
+  "version": "3.2.2",
   "description": "Autonomous AI coding loop - PRD-driven development with Claude Code",
   "author": "Allie Jones <allie@allthrive.ai>",
   "license": "MIT",

package/ralph/setup.sh

@@ -40,8 +40,7 @@ ralph_setup() {
   echo "  /idea 'your feature'    # Generate a PRD"
   echo ""
   echo "  --- Terminal 2: Ralph Loop ---"
-  echo "  npx agentic-loop run       # Execute PRDs autonomously"
-  echo "  npx agentic-loop run --fast  # Skip code review (~2x faster)"
+  echo "  npx agentic-loop run         # Execute PRDs autonomously"
   echo ""
 }
 
@@ -359,7 +358,7 @@ setup_mcp() {
     tmp=$(mktemp)
     jq '.mcpServers["playwright"] = {
       "command": "npx",
-      "args": ["-y", "@anthropic-ai/mcp-server-playwright"]
+      "args": ["-y", "@playwright/mcp@latest"]
     }' "$claude_json" > "$tmp" && mv "$tmp" "$claude_json"
     echo "  Added playwright MCP server (browser automation & testing)"
     added_any=true
@@ -407,20 +406,18 @@ setup_test_credentials() {
     read -r -s -p "  Test user password: " test_password
     echo ""
 
-    # Add to .env (append or update)
+    # Add to .env (remove old entries first, then append)
     if grep -q "^RALPH_TEST_USER=" .env 2>/dev/null; then
-      # Update existing
+      # Remove old entries
       local tmp
       tmp=$(mktemp)
-      sed "s/^RALPH_TEST_USER=.*/RALPH_TEST_USER=$test_email/" .env > "$tmp" && mv "$tmp" .env
-      sed "s/^RALPH_TEST_PASSWORD=.*/RALPH_TEST_PASSWORD=$test_password/" .env > "$tmp" && mv "$tmp" .env
-    else
-      # Append new
-      echo "" >> .env
-      echo "# Test credentials for browser automation (auto-added by agentic-loop)" >> .env
-      echo "RALPH_TEST_USER=$test_email" >> .env
-      echo "RALPH_TEST_PASSWORD=$test_password" >> .env
+      grep -v "^RALPH_TEST_USER=\|^RALPH_TEST_PASSWORD=" .env > "$tmp" && mv "$tmp" .env
     fi
+    # Append new (using printf to handle special characters safely)
+    echo "" >> .env
+    echo "# Test credentials for browser automation (auto-added by agentic-loop)" >> .env
+    printf 'RALPH_TEST_USER=%s\n' "$test_email" >> .env
+    printf 'RALPH_TEST_PASSWORD=%s\n' "$test_password" >> .env
 
     echo "  Saved credentials to .env"
     echo ""

package/templates/PROMPT.md

@@ -31,15 +31,14 @@ For each story, you must:
 **You have browser tools - USE THEM to verify your work:**
 
 **Playwright MCP** (testing & automation):
-- `browser_navigate` - Go to a URL and get page content
-- `browser_screenshot` - Take a screenshot to verify UI
-- `browser_click` - Click elements to test interactions
-- `browser_type` - Fill in forms to test inputs
-- `browser_snapshot` - Get accessibility tree for a11y testing
+- Navigate to URLs and verify page content
+- Take screenshots to verify UI renders correctly
+- Click elements and fill forms to test interactions
+- Get accessibility snapshots for a11y testing
 
 **Chrome DevTools MCP** (debugging & inspection):
-- Inspect DOM, check console for errors
-- Debug network requests
+- Inspect DOM elements and check console for errors
+- Debug network requests and responses
 - Check element styles and computed properties
 
 **Do NOT say you're done until:**