diff --git a/kalibr/cli/prompt_cmd.py b/kalibr/cli/prompt_cmd.py
index 0e29d4e..59d3c51 100644
--- a/kalibr/cli/prompt_cmd.py
+++ b/kalibr/cli/prompt_cmd.py
@@ -22,14 +22,20 @@
 
 1. pip install kalibr huggingface_hub
 
-2. kalibr auth --agent --email <YOUR_EMAIL>
+2. kalibr signup <YOUR_EMAIL>
+   (Creates your Kalibr account. Credentials saved to .env immediately.)
 
 3. Source credentials into shell:
-   - If ~/.zshrc exists: append KALIBR_API_KEY and KALIBR_TENANT_ID exports
-   - Else if ~/.bashrc exists: append to ~/.bashrc
-   - Run: source ~/.zshrc (or ~/.bashrc)
+   export $(grep -v '^#' .env | xargs)
+   To persist across sessions, append to ~/.zshrc:
+     grep 'KALIBR_API_KEY\\|KALIBR_TENANT_ID' .env >> ~/.zshrc
+     source ~/.zshrc
 
-4. openclaw plugin install kalibr && openclaw plugin configure kalibr
+4. openclaw plugins install @kalibr/openclaw
+   openclaw config set plugins.entries.kalibr.config.apiKey "$KALIBR_API_KEY"
+   openclaw config set plugins.entries.kalibr.config.tenantId "$KALIBR_TENANT_ID"
+   openclaw config set plugins.entries.kalibr.config.enableRouting true
+   openclaw gateway restart
 
 5. Create ~/clawd workspace scaffold:
    - ~/clawd/SOUL.md — ask the user: "What does your agent do in one sentence?"
@@ -41,11 +47,62 @@
    - ~/clawd/AGENTS.md — write a routing table:
      # AGENTS.md\\n\\n## Routing Table\\n\\n| Task Type | Load | Route |\\n|-----------|------|-------|\\n| code_generation / code_review / system_design | high | Sonnet → GPT-4o |\\n| research / outreach_generation | medium | Llama → DeepSeek |\\n| web_scraping / classification / summarization / data_pipeline | low | DeepSeek → Llama |
 
-6. Create ~/clawd/kalibr-integration/ directory and download pipeline files:
-   - https://raw.githubusercontent.com/kalibr-ai/kalibr-sdk-python/main/examples/kalibr_pipeline.py
-   - https://raw.githubusercontent.com/kalibr-ai/kalibr-sdk-python/main/examples/kalibr_taxonomy.py
-   - https://raw.githubusercontent.com/kalibr-ai/kalibr-sdk-python/main/examples/kalibr_eval.py
-   - Create an empty ~/clawd/kalibr-integration/__init__.py
+6. Create ~/clawd/kalibr-integration/ directory and write these files:
+
+   Create ~/clawd/kalibr-integration/__init__.py (empty file)
+
+   Create ~/clawd/kalibr-integration/kalibr_pipeline.py with this content:
+   """
+OPENCLAW_PIPELINE_FILE = '''"""Kalibr pipeline integration — classify and route tasks."""
+import os
+from kalibr import Router
+
+
+def classify_and_route(task: str, context: dict = None) -> dict:
+    """Classify a task and route it through Kalibr."""
+    goal_id = classify_task(task)
+    router = Router(goal=goal_id, paths=get_paths_for_goal(goal_id))
+    result = router.completion(messages=[{"role": "user", "content": task}])
+    return {"goal": goal_id, "result": result.choices[0].message.content}
+
+
+def classify_task(task: str) -> str:
+    """Classify a task string to a Kalibr goal_id."""
+    t = task.lower()
+    if any(w in t for w in ["scrape", "extract from url", "fetch page"]): return "web_scraping"
+    if any(w in t for w in ["enrich", "fill in", "lookup missing"]): return "data_enrichment"
+    if any(w in t for w in ["score lead", "lead score", "rate prospect"]): return "lead_scoring"
+    if any(w in t for w in ["classify", "categorize", "label"]): return "classification"
+    if any(w in t for w in ["summarize", "summary", "tldr"]): return "summarization"
+    if any(w in t for w in ["pipeline", "transform data", "process rows"]): return "data_pipeline"
+    if any(w in t for w in ["research", "investigate", "analyze market"]): return "research"
+    if any(w in t for w in ["draft email", "cold email", "outreach", "write message"]): return "outreach_generation"
+    if any(w in t for w in ["write code", "implement", "build function", "create script"]): return "code_generation"
+    if any(w in t for w in ["review code", "code review", "check this code"]): return "code_review"
+    if any(w in t for w in ["design system", "architecture", "system design"]): return "system_design"
+    return "research"
+
+
+def get_paths_for_goal(goal_id: str) -> list:
+    """Return default model paths for a goal_id."""
+    paths = {
+        "web_scraping": ["deepseek-chat", "meta-llama/Llama-3.3-70B-Instruct", "mistralai/Mixtral-8x7B-Instruct-v0.1", "gpt-4o-mini"],
+        "data_enrichment": ["deepseek-chat", "meta-llama/Llama-3.3-70B-Instruct", "Qwen/Qwen2.5-72B-Instruct", "gpt-4o-mini"],
+        "lead_scoring": ["deepseek-chat", "meta-llama/Llama-3.3-70B-Instruct", "mistralai/Mixtral-8x7B-Instruct-v0.1", "gpt-4o-mini"],
+        "classification": ["deepseek-chat", "meta-llama/Llama-3.3-70B-Instruct", "Qwen/Qwen2.5-72B-Instruct", "gpt-4o-mini"],
+        "summarization": ["deepseek-chat", "meta-llama/Llama-3.3-70B-Instruct", "mistralai/Mixtral-8x7B-Instruct-v0.1", "claude-haiku-3-5-20241022"],
+        "data_pipeline": ["deepseek-chat", "meta-llama/Llama-3.3-70B-Instruct", "Qwen/Qwen2.5-72B-Instruct", "gpt-4o-mini"],
+        "research": ["meta-llama/Llama-3.3-70B-Instruct", "deepseek-chat", "deepseek-reasoner", "claude-sonnet-4-20250514"],
+        "outreach_generation": ["meta-llama/Llama-3.3-70B-Instruct", "deepseek-chat", "mistralai/Mixtral-8x22B-Instruct-v0.1", "claude-sonnet-4-20250514"],
+        "code_generation": ["claude-sonnet-4-20250514", "gpt-4o", "o3-mini", "deepseek-reasoner"],
+        "code_review": ["claude-sonnet-4-20250514", "gpt-4o", "deepseek-reasoner", "o3-mini"],
+        "system_design": ["claude-sonnet-4-20250514", "deepseek-reasoner", "gpt-4o", "o3-mini"],
+        "agent_orchestration": ["claude-sonnet-4-20250514", "gpt-4o", "deepseek-reasoner", "o3-mini"],
+    }
+    return paths.get(goal_id, paths["research"])
+'''
+
+OPENCLAW_PROMPT_TEXT_PART2 = """\
 
 7. kalibr init && kalibr verify
 
@@ -57,11 +114,17 @@
 9. Report done:
    - Print: "✅ Kalibr + OpenClaw setup complete."
    - Print: "Dashboard: https://dashboard.kalibr.systems"
-   - Print the claim email confirmation URL returned by `kalibr auth --agent`
+   - Print the claim_url returned by kalibr signup (the human uses it to access the dashboard)
    - Ask: "Do you have any provider API keys to add? (OPENAI_API_KEY, ANTHROPIC_API_KEY, DEEPSEEK_API_KEY, HF_TOKEN)"
      If yes, append them to ~/.zshrc and source it.
 """
 
+OPENCLAW_FULL_PROMPT = (
+    OPENCLAW_PROMPT_TEXT
+    + OPENCLAW_PIPELINE_FILE
+    + OPENCLAW_PROMPT_TEXT_PART2
+)
+
 
 def prompt(
     openclaw: bool = typer.Option(
@@ -82,12 +145,12 @@ def prompt(
     """
     if openclaw:
         if email:
-            text = OPENCLAW_PROMPT_TEXT.replace(
-                "kalibr auth --agent --email <YOUR_EMAIL>",
-                f"kalibr auth --agent --email {email}",
+            text = OPENCLAW_FULL_PROMPT.replace(
+                "kalibr signup <YOUR_EMAIL>",
+                f"kalibr signup {email}",
             )
         else:
-            text = OPENCLAW_PROMPT_TEXT
+            text = OPENCLAW_FULL_PROMPT
     else:
         text = PROMPT_TEXT
 
diff --git a/pyproject.toml b/pyproject.toml
index bdbd686..faa833d 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "kalibr"
-version = "1.9.7"
+version = "1.9.8"
 description = "Outcome-aware LLM routing for production AI agents. Routes between models, tools, and parameters based on real success signals using Thompson Sampling. Automatic fallback, cost optimization, and continuous learning — no redeploy required."
 authors = [{name = "Kalibr Team", email = "support@kalibr.systems"}]
 readme = "README.md"