import requests
from dotenv import load_dotenv
import os
import json
load_dotenv()
api_key = os.getenv("LANGSMITH_API_KEY")
BASE_URL = "https://api.smith.langchain.com" # adjust if using a different host
def auth_headers():
return {
"X-API-Key": api_key
}
def list_rules(params=None):
url = f"{BASE_URL}/api/v1/runs/rules"
resp = requests.get(url, headers=auth_headers(), params=params or {}, timeout=30)
if resp.status_code != 200:
raise RuntimeError(f"List rules failed: {resp.status_code} {resp.text}")
return resp.json()
def get_rule_by_id(rule_id: str):
"""
The API doesn't provide GET /rules/{id}, so we filter the list endpoint by id.
The 'id' parameter accepts an array of strings, so we pass [rule_id].
"""
rules = list_rules(params={"id": [rule_id]})
if not rules:
print(f"No rule found with id: {rule_id}")
return None
rule = rules[0]
print("Rule details:\n" + json.dumps(rule, indent=2))
return rule
def create_prompt_injection_clone():
url = f"{BASE_URL}/api/v1/runs/rules"
payload = {
"display_name": "prompt-injection-2", # new name
"session_id": "503e8389-f655-4e5b-90cf-548af7da9875",
"is_enabled": True,
"dataset_id": None,
"sampling_rate": 1.0,
"filter": "eq(is_root, true)",
"trace_filter": None,
"tree_filter": None,
"use_corrections_dataset": False,
"num_few_shot_examples": None,
"extend_only": False,
"transient": False,
"add_to_annotation_queue_id": None,
"add_to_dataset_id": None,
"add_to_dataset_prefer_correction": False,
"evaluators": [
{
"structured": {
"hub_ref": "eval_chat_langchain_v3_prompt_injection_fda713cd:latest",
"prompt": None,
"template_format": None,
"schema": None,
"variable_mapping": {
"input": "input.messages"
},
"model": None
}
}
],
"code_evaluators": [],
"alerts": [],
"webhooks": [],
"evaluator_version": 3,
"create_alignment_queue": False,
"include_extended_stats": False
}
resp = requests.post(url, headers=auth_headers(), json=payload, timeout=30)
if resp.status_code == 200:
print("✅ Rule created successfully:")
print(json.dumps(resp.json(), indent=2))
else:
print(f"❌ Error {resp.status_code}: {resp.text}")
if __name__ == "__main__":
print("=== All rules (id :: name) ===")
params = {
"name_contains": "prompt-injection"
}
for r in list_rules(params=params):
print(f"{r['id']} :: {r.get('display_name')}")
rule_id = r['id']
print("\n=== Selected rule ===")
get_rule_by_id(rule_id)
print("\n=== Creating prompt-injection-2 ===")
# create_prompt_injection_clone()
