Integrations

SpiderShield works with every major AI agent framework. Add security in minutes.

LangChain

Stable

Wrap LangChain tool calls with SpiderGuard for policy enforcement and DLP scanning.

python
from spidershield import SpiderGuard
from langchain.tools import Tool

guard = SpiderGuard(policy="balanced", dlp="redact")

# Wrap any LangChain tool
@guard.langchain_wrapper
def my_tool(query: str) -> str:
    return db.execute(query)

OpenAI Agents

Stable

Guard OpenAI function calls with pre/post execution checks and audit logging.

python
from spidershield import SpiderGuard

guard = SpiderGuard(policy="strict")

# Before function execution
result = guard.check(fn_name, fn_args)
if result.decision == Decision.ALLOW:
    output = execute_function(fn_name, fn_args)
    clean = guard.after_check(fn_name, output)

CrewAI

Stable

Add security guardrails to CrewAI agent crews with zero config changes.

python
from spidershield import SpiderGuard

guard = SpiderGuard(policy="balanced")

# Guard CrewAI tool execution
result = guard.check("search_web",
    {"query": user_input})
# Decision: ALLOW / DENY / ESCALATE

AutoGen

Stable

Intercept AutoGen agent tool calls with policy-based access control.

python
from spidershield import SpiderGuard

guard = SpiderGuard(policy="balanced")

# In your AutoGen agent's tool handler
def guarded_tool(name, args):
    check = guard.check(name, args)
    if check.decision == Decision.DENY:
        return f"Blocked: {check.reason}"
    return original_tool(name, args)

MCP Servers

Stable

Zero-code proxy mode for any MCP server. No SDK changes required.

bash
# Wrap any MCP server with SpiderShield proxy
"code-keyword">$ "code-function">spidershield proxy \
    "code-keyword">--policy strict \
    -- python my_mcp_server.py

# Or use guard mode for stdio servers
"code-keyword">$ "code-function">spidershield guard \
    "code-keyword">--preset balanced \
    -- "code-function">npx @modelcontextprotocol/server-filesystem /tmp

Claude Desktop

Beta

Add SpiderShield as an MCP proxy in your Claude Desktop config for automatic protection.

bash
// claude_desktop_config.json
{
  "mcpServers": {
    "filesystem-guarded": {
      "command": "">spidershield",
      "args": ["guard", "">--preset", "balanced",
               "--", "">npx",
               "@modelcontextprotocol/server-filesystem",
               "/home/user/safe-dir"]
    }
  }
}

Don't see your framework? SpiderShield's Python SDK works with any tool-calling agent.

Read the DocsRequest Integration