> ## Documentation Index
> Fetch the complete documentation index at: https://docs.bevor.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Creating Agents & Tools

> Building custom security agents and workflows

# Creating Agents & Tools

Build specialized security agents and custom workflows tailored to your specific protocol needs. BevorAI's agent framework enables you to create intelligent, autonomous security tools that understand your unique requirements and risk profile.

<Note>
  If your Bevor-powered agents win competitive audits against human auditors on platforms like Sherlock, Code4rena, or Cantina, you may be eligible for additional rewards! Contact us with proof or a request to sponsor your project at [contact@bevor.io](mailto:contact@bevor.io)
</Note>

## Popular Agent Frameworks

BevorAI can be integrated into leading agent frameworks to create specialized security monitoring and analysis tools.

<CardGroup cols={2}>
  <Card title="LangChain" icon="link-simple" href="https://python.langchain.com/docs/get_started/introduction">
    Build agents that combine BevorAI security analysis with LangChain's powerful orchestration
  </Card>

  <Card title="AutoGPT" icon="robot" href="https://github.com/Significant-Gravitas/AutoGPT">
    Create autonomous security agents that proactively monitor and respond to threats
  </Card>

  <Card title="BabyAGI" icon="brain" href="https://github.com/yoheinakajima/babyagi">
    Develop task-driven security workflows with recursive improvement
  </Card>

  <Card title="Agents.js" icon="js" href="https://github.com/agents-js/agents.js">
    JavaScript-native agents for web3-focused security automation
  </Card>
</CardGroup>

## Integration Patterns

### Basic API Integration

#### Requirements

* Feed compilable code to the agent. Provide sources (upload/repo) or a deployed address with verified sources so Bevor can compile and analyze.
* Include compiler metadata when available (e.g., Solidity `solc` version). See <a href="/api-reference/code-versions">Code Versions API</a>.

#### Minimal flow

1. Create a code version (upload or scan by address)
2. Create an audit for that version
3. Poll audit status until complete
4. Fetch findings and act

```bash theme={null}
# 1) Create code version from a verified contract address
curl -sS -X POST https://api.bevor.io/v1/code-versions \
  -H "Authorization: Bearer $BEVOR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"version_method":"address","version_identifier":"0xYourContract","source_type":"scan"}'

# 2) Create an audit for that code version
curl -sS -X POST https://api.bevor.io/v1/audits \
  -H "Authorization: Bearer $BEVOR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"version_mapping_id":"<VERSION_MAPPING_ID>"}'

# 3/4) Poll status and fetch results
curl -sS -H "Authorization: Bearer $BEVOR_API_KEY" \
  https://api.bevor.io/v1/audits/<AUDIT_ID>
```

Refer to <a href="/api-reference/audits">Audits API</a> for details.

### Granular integrations via Chat

* After an audit completes, your agent can ask targeted questions, request explanations, or generate actions using the Chat endpoints (great for triage or reporting flows).
* Use Chat to enrich tickets, summarize findings, or request remediation steps per finding.
* See endpoints in the <a href="/api-reference/introduction">API Reference</a>.

## Next Steps

<CardGroup cols={2}>
  <Card title="Protocol Integration" icon="puzzle-piece" href="/developers/protocol-integration">
    Learn how to integrate BevorAI into your protocol
  </Card>

  <Card title="API Reference" icon="code" href="/api-reference/introduction">
    Explore the complete API documentation
  </Card>
</CardGroup>