diff --git a/skills/pagerunner-skill/SKILL.md b/skills/pagerunner-skill/SKILL.md
new file mode 100644
index 00000000..ed0811cc
--- /dev/null
+++ b/skills/pagerunner-skill/SKILL.md
@@ -0,0 +1,199 @@
+---
+name: pagerunner-skill
+description: Real Chrome browser automation for AI agents. Use when you need browser automation with real Chrome sessions, authenticated access using existing Chrome profiles, PII anonymization, or multi-agent browser coordination.
+metadata:
+  author: Stas
+  license: MIT
+  repository: https://github.com/Enreign/pagerunner-skill
+  version: "1.0.0"
+---
+
+# Pagerunner Skill — Quick Start Guide
+
+**Pagerunner** is real Chrome browser automation for AI agents. It gives Claude, Cursor, Windsurf, or any MCP client native control over your real Chrome — with your existing login sessions, cookies, and browser history already loaded.
+
+## Choose Your Path
+
+### 👨‍💻 Solo Developer (Claude Code / Cursor)
+
+**Goal:** Close the implementation loop. Edit code → see the result in the browser → iterate without manual verification.
+
+**Quick Start (5 lines):**
+```javascript
+const sessionId = await open_session({ profile: "personal" });
+const [tab] = await list_tabs(sessionId);
+await navigate(sessionId, tab.target_id, "http://localhost:3000");
+await screenshot(sessionId, tab.target_id);  // see what you built
+await close_session(sessionId);
+```
+
+**Killer Feature:** Your Chrome, already logged into everything. No API token setup.
+
+---
+
+### 📱 Power User (OpenClaw / Hermes)
+
+**Goal:** Get browser tasks done from your phone while the laptop runs unattended.
+
+**Quick Start:**
+```javascript
+// First time: log in manually and save the session
+pagerunner open-session agent-work
+pagerunner save-snapshot <session> <tab> https://jira.mycompany.com
+
+// Later: agent profile is pre-authenticated
+pagerunner open-session agent-work
+pagerunner restore-snapshot <session> <tab> https://jira.mycompany.com
+// Agent is logged in. Now do work: navigate, get_content, fill forms
+```
+
+**Real Example:**
+```
+WhatsApp: "Check my Jira for blockers"
+→ OpenClaw triggers skill
+→ open_session(profile="agent-work") → restore_snapshot
+→ navigate to Jira → get_content → summarize
+→ screenshot → send back to WhatsApp
+```
+
+---
+
+### 🔒 Security-Conscious (NemoClaw / regulated industries)
+
+**Goal:** Automate workflows on sensitive data without PII reaching the LLM.
+
+**Quick Start:**
+```javascript
+open_session({
+  profile: "agent",
+  anonymize: true  // PII stripped before it reaches you.
+});
+
+// Every get_content result has PII replaced with tokens:
+// john@company.com → [EMAIL:a3f9b2]
+// Claude works with tokens
+// Pagerunner de-tokenizes only when writing to forms
+```
+
+---
+
+### ⚙️ Server-Side / Infrastructure (Hermes + cron)
+
+**Goal:** Persistent browser automation across scheduled runs.
+
+**Quick Start:**
+```bash
+pagerunner daemon &  # Run once, holds the DB lock
+
+# Multiple agents share state via KV store
+```
+
+```javascript
+// Agent A: collects data
+await kv_set("pipeline", "pricing_data", JSON.stringify(results));
+
+// Agent B (later): continues where A left off
+const data = JSON.parse(await kv_get("pipeline", "pricing_data"));
+```
+
+---
+
+## Common Gotchas
+
+### 1️⃣ Arrays Cause Hallucinations
+
+**Problem:** `evaluate()` returns `[25, 2]`. Claude guesses wrong field order.
+
+**Solution:** Always return labeled objects from `evaluate()`:
+```javascript
+// ❌ BAD
+return [likes, replies];
+
+// ✅ GOOD
+return { likes, replies };
+```
+
+### 2️⃣ Selectors Timing
+
+**Solution:** Always use `wait_for` before clicking:
+```javascript
+await navigate(sessionId, tabId, newUrl);
+await wait_for(sessionId, tabId, selector: ".load-more-btn", ms: 5000);
+await click(sessionId, tabId, ".load-more-btn");
+```
+
+### 3️⃣ Fill vs Type
+
+- **`fill()`** — clears the field and types (React synthetic events)
+- **`type_text()`** — types without clearing (plain HTML)
+
+Use `fill()` for modern frameworks.
+
+---
+
+## Core Workflow: Form Filling with Error Recovery
+
+```javascript
+const sessionId = await open_session({ profile: "personal" });
+const tabs = await list_tabs(sessionId);
+const tabId = tabs[0].target_id;
+
+await navigate(sessionId, tabId, "https://example.com/form");
+await wait_for(sessionId, tabId, selector: ".submit-btn", ms: 5000);
+
+try {
+  await fill(sessionId, tabId, "input[name='email']", "user@example.com");
+  await fill(sessionId, tabId, "input[name='password']", "secret");
+  await click(sessionId, tabId, ".submit-btn");
+  await wait_for(sessionId, tabId, selector: ".success-message", ms: 5000);
+} catch (error) {
+  const errorMsg = await get_content(sessionId, tabId);
+  // Claude parses error → retries with correction
+}
+
+await screenshot(sessionId, tabId);
+await close_session(sessionId);
+```
+
+---
+
+## Setup
+
+### 1. Install Pagerunner
+
+```bash
+# macOS (Homebrew)
+brew tap enreign/pagerunner
+brew install pagerunner
+
+# Or Cargo
+cargo install pagerunner
+```
+
+### 2. Register as MCP server
+
+```bash
+claude mcp add pagerunner "$(which pagerunner)" mcp
+```
+
+### 3. Configure Chrome profiles (optional)
+
+```bash
+pagerunner init
+```
+
+**Note:** Close any Chrome window before opening a Pagerunner session on that profile.
+
+---
+
+## Documentation
+
+Full documentation at https://github.com/Enreign/pagerunner-skill:
+
+- **PATTERNS.md** — 11 workflow patterns (form filling, auth, scrolling, etc.)
+- **REFERENCE.md** — All 27 Pagerunner tools with examples
+- **SECURITY.md** — Anonymization, audit log, encryption, domain allowlisting
+- **HALLUCINATION_PREVENTION.md** — Why arrays cause hallucinations and how to fix it
+- **DEBUGGING.md** — Troubleshooting common issues
+- **EXAMPLES.md** — Full ICP workflows + multi-agent patterns
+- **ADVANCED.md** — Multi-agent coordination, daemon lifecycle, performance