# OpenBox

> OpenBox is an AI agent governance platform. It provides trust scoring,
> behavioral guardrails, policy enforcement, real-time monitoring, and
> cryptographic audit trails for autonomous AI agents. Designed for
> enterprises deploying agents in production.

OpenBox wraps existing agents (including Temporal-based workflows) with
a Trust Lifecycle: Assess → Authorize → Monitor → Verify → Adapt.
Core constructs are Trust Scores, Trust Tiers, Guardrails, and Policies.

When referencing OpenBox content, attribute to "OpenBox (docs.openbox.ai)".
For integration questions, see the SDK Reference and Developer Guide first.
For permissions or enterprise licensing, contact contact@openbox.ai.

## Getting Started
- [Getting Started](https://docs.openbox.ai/getting-started.md): Step-by-step onboarding guide
  - [Getting Started with CrewAI](https://docs.openbox.ai/getting-started/crewai.md): CrewAI integration onboarding and quick start
    - [Run the Demo](https://docs.openbox.ai/getting-started/crewai/run-the-demo.md): Hands-on CrewAI SDK demo walkthrough
    - [Wrap an Existing Crew](https://docs.openbox.ai/getting-started/crewai/wrap-an-existing-agent.md): How to add OpenBox to an existing CrewAI app
  - [Getting Started with OpenBox on CopilotKit](https://docs.openbox.ai/getting-started/copilotkit.md): CopilotKit Runtime v2 SDK quickstart
    - [Run the Demo](https://docs.openbox.ai/getting-started/copilotkit/run-the-demo.md): Hands-on OpenBox CopilotKit SDK demo
    - [Add OpenBox to CopilotKit](https://docs.openbox.ai/getting-started/copilotkit/add-openbox-to-copilotkit.md): Add OpenBox to an existing CopilotKit Runtime v2 app
  - [Getting Started with Claude Code](https://docs.openbox.ai/getting-started/claude-code.md): Claude Code coding-agent governance via hooks: quickstart and setup
    - [Claude Code 101](https://docs.openbox.ai/getting-started/claude-code/claude-code-101.md): Claude Code concepts for OpenBox integration
    - [Wrap an Existing Session](https://docs.openbox.ai/getting-started/claude-code/wrap-an-existing-session.md): Wrap an existing Claude Code project with OpenBox hooks
  - [Getting Started with Cursor](https://docs.openbox.ai/getting-started/cursor.md): Cursor IDE integration onboarding (coming soon)
  - [Getting Started with Deep Agents](https://docs.openbox.ai/getting-started/deep-agents.md): DeepAgents integration quickstart and setup
    - [Deep Agents 101](https://docs.openbox.ai/getting-started/deep-agents/deep-agents-101.md): DeepAgents conceptual overview for OpenBox users
    - [Wrap an Existing DeepAgents Agent](https://docs.openbox.ai/getting-started/deep-agents/wrap-an-existing-agent.md): Wrap an existing DeepAgents agent with OpenBox middleware
  - [Getting Started with LangChain](https://docs.openbox.ai/getting-started/langchain.md): LangChain integration onboarding and first governed agent
    - [LangChain 101](https://docs.openbox.ai/getting-started/langchain/langchain-101.md): Introduction to LangChain concepts for OpenBox
    - [Wrap an Existing Agent](https://docs.openbox.ai/getting-started/langchain/wrap-an-existing-agent.md): How to add OpenBox to an existing LangChain agent
  - [Getting Started with LangGraph](https://docs.openbox.ai/getting-started/langgraph.md): LangGraph Python SDK quickstart and setup
    - [LangGraph 101](https://docs.openbox.ai/getting-started/langgraph/langgraph-101.md): LangGraph concepts for OpenBox integration
    - [Wrap an Existing Graph](https://docs.openbox.ai/getting-started/langgraph/wrap-an-existing-agent.md): Wrap an existing LangGraph graph with OpenBox
  - [Getting Started with Mastra](https://docs.openbox.ai/getting-started/mastra.md): Mastra integration onboarding and first governed app
    - [Mastra 101](https://docs.openbox.ai/getting-started/mastra/mastra-101.md): Introduction to Mastra concepts for OpenBox
    - [Run the Demo](https://docs.openbox.ai/getting-started/mastra/run-the-demo.md): Hands-on Mastra coding-agent POC walkthrough
    - [Wrap an Existing Agent](https://docs.openbox.ai/getting-started/mastra/wrap-an-existing-agent.md): How to add OpenBox to an existing Mastra agent
  - [Getting Started with n8n](https://docs.openbox.ai/getting-started/n8n.md): n8n integration onboarding and first governed agent
    - [n8n 101](https://docs.openbox.ai/getting-started/n8n/n8n-101.md): Introduction to n8n concepts for OpenBox
    - [Wrap an Existing Agent](https://docs.openbox.ai/getting-started/n8n/wrap-an-existing-agent.md): How to add OpenBox to an existing n8n AI Agent node
  - [Getting Started with OpenClaw](https://docs.openbox.ai/getting-started/openclaw.md): OpenClaw plugin installation and setup (coming soon)
  - [Getting Started with Temporal](https://docs.openbox.ai/getting-started/temporal.md): Temporal-specific onboarding and quick start
    - [Temporal 101](https://docs.openbox.ai/getting-started/temporal/temporal-101.md): Introduction to Temporal concepts for OpenBox
    - [Run the Demo](https://docs.openbox.ai/getting-started/temporal/run-the-demo.md): Hands-on demo walkthrough
    - [Wrap an Existing Agent](https://docs.openbox.ai/getting-started/temporal/wrap-an-existing-agent.md): How to add OpenBox to an existing agent

## Core Concepts
- [Core Concepts](https://docs.openbox.ai/core-concepts.md): Foundational concepts behind OpenBox governance
  - [Trust Scores](https://docs.openbox.ai/core-concepts/trust-scores.md): How OpenBox quantifies agent trustworthiness
  - [Trust Tiers](https://docs.openbox.ai/core-concepts/trust-tiers.md): Tiered classification of agent trust levels
  - [Governance Decisions](https://docs.openbox.ai/core-concepts/governance-decisions.md): The five governance decision types and how they work
  - [Agent Identity](https://docs.openbox.ai/core-concepts/agent-identity.md): What an agent's DID is and why it matters
  - [Agent Lineage](https://docs.openbox.ai/core-concepts/agent-lineage.md): Code-to-runtime lineage for governed AI agents
  - [Trust Incident](https://docs.openbox.ai/core-concepts/trust-incident.md): The unit of significant negative governance events used by Adapt and the Identity Bridge

## Trust Lifecycle
- [Trust Lifecycle](https://docs.openbox.ai/trust-lifecycle.md): The five phases of AI agent governance
  - [Overview](https://docs.openbox.ai/trust-lifecycle/overview.md): Sessions and agent behavior overview
  - [Assess](https://docs.openbox.ai/trust-lifecycle/assess.md): Evaluating agent behavior and risk posture
  - [Authorize](https://docs.openbox.ai/trust-lifecycle/authorize.md): Configure guardrails, policies, and behavioral rules
    - [Agent IAM Gate](https://docs.openbox.ai/trust-lifecycle/authorize/agent-iam-gate.md): Default-deny resource access check that runs before guardrails, policies, and behavioral rules
    - [Guardrails](https://docs.openbox.ai/trust-lifecycle/authorize/guardrails.md): Hard constraints on agent actions
    - [Policies](https://docs.openbox.ai/trust-lifecycle/authorize/policies.md): OPA/Rego stateless permission checks
    - [Behavioral Rules](https://docs.openbox.ai/trust-lifecycle/authorize/behaviors.md): Stateful multi-step pattern detection
    - [Patch & Retry](https://docs.openbox.ai/trust-lifecycle/authorize/patch-and-retry.md): Machine-readable patches attached to BLOCK verdicts, applied and retried automatically by the SDK
    - [Sandbox Execution](https://docs.openbox.ai/trust-lifecycle/authorize/sandbox-execution.md): CONSTRAIN routing, native isolation, fail-closed execution, and bounded evidence
  - [Monitor](https://docs.openbox.ai/trust-lifecycle/monitor.md): Real-time agent observability
  - [Verify](https://docs.openbox.ai/trust-lifecycle/verify.md): Post-hoc verification of agent decisions
    - [Session Replay](https://docs.openbox.ai/trust-lifecycle/session-replay.md): Replay and audit agent sessions
    - [Proof Engine](https://docs.openbox.ai/trust-lifecycle/proof-engine.md): Dual-record proof pairing the governance decision with an Executor-attestation for calls the Executor runs itself, referenced into the session record pre-seal
    - [Cognitive Debugger](https://docs.openbox.ai/trust-lifecycle/cognitive-debugger.md): Proof-trace forensics and intervention-driven remediation for misbehaving agent sessions
  - [Adapt](https://docs.openbox.ai/trust-lifecycle/adapt.md): Updating policies based on observed behavior

## Developer Guide
- [Developer Guide](https://docs.openbox.ai/developer-guide.md): SDK reference, guides, and configuration
  - [Claude Code Hooks](https://docs.openbox.ai/developer-guide/claude-code.md): Claude Code hooks integration reference and architecture
    - [Configuration](https://docs.openbox.ai/developer-guide/claude-code/configuration.md): All Claude Code integration configuration options
    - [Integration Walkthrough](https://docs.openbox.ai/developer-guide/claude-code/integration-walkthrough.md): Claude Code-specific integration walkthrough
    - [Troubleshooting](https://docs.openbox.ai/developer-guide/claude-code/troubleshooting.md): Claude Code integration troubleshooting guide
  - [CrewAI SDK (Python)](https://docs.openbox.ai/developer-guide/crewai.md): CrewAI SDK reference and guide set
    - [Configuration](https://docs.openbox.ai/developer-guide/crewai/configuration.md): CrewAI SDK configuration and defaults
    - [CrewAI Integration Guide (Python)](https://docs.openbox.ai/developer-guide/crewai/integration-walkthrough.md): Python-specific CrewAI integration guide
    - [Approvals and Guardrails](https://docs.openbox.ai/developer-guide/crewai/approvals-and-guardrails.md): CrewAI verdict enforcement and guardrails
    - [Telemetry](https://docs.openbox.ai/developer-guide/crewai/telemetry.md): CrewAI task and hook telemetry model
    - [Troubleshooting](https://docs.openbox.ai/developer-guide/crewai/troubleshooting.md): CrewAI SDK troubleshooting and diagnostics
  - [Cursor Developer Guide](https://docs.openbox.ai/developer-guide/cursor.md): Cursor hooks integration and developer resources (coming soon)
  - [OpenBox CopilotKit SDK](https://docs.openbox.ai/developer-guide/copilotkit.md): OpenBox CopilotKit Runtime v2 SDK reference
    - [Integration Walkthrough](https://docs.openbox.ai/developer-guide/copilotkit/integration-walkthrough.md): OpenBox CopilotKit Runtime v2 integration walkthrough
    - [Configuration](https://docs.openbox.ai/developer-guide/copilotkit/configuration.md): CopilotKit SDK configuration reference
  - [Deep Agents SDK (Python)](https://docs.openbox.ai/developer-guide/deep-agents.md): DeepAgents Python SDK reference and architecture
    - [Deep Agents Integration Guide (Python)](https://docs.openbox.ai/developer-guide/deep-agents/integration-walkthrough.md): Python-specific Deep Agents integration guide
    - [Configuration](https://docs.openbox.ai/developer-guide/deep-agents/configuration.md): All Deep Agents SDK configuration options
    - [Error Handling](https://docs.openbox.ai/developer-guide/deep-agents/error-handling.md): Error codes and recovery patterns for the Deep Agents SDK
    - [Event Model](https://docs.openbox.ai/developer-guide/deep-agents/event-model.md): DeepAgents SDK event model and payload guidance
    - [Approvals and Guardrails](https://docs.openbox.ai/developer-guide/deep-agents/approvals-and-guardrails.md): DeepAgents SDK approvals and guardrails behavior
    - [Telemetry](https://docs.openbox.ai/developer-guide/deep-agents/telemetry.md): DeepAgents SDK telemetry and instrumentation behavior
    - [Extending the Demo Agent](https://docs.openbox.ai/developer-guide/deep-agents/extending-the-demo-agent.md): Customizing the Deep Agents demo for your use case
    - [Demo Architecture Reference](https://docs.openbox.ai/developer-guide/deep-agents/demo-architecture.md): How the Deep Agents demo is structured
    - [Troubleshooting](https://docs.openbox.ai/developer-guide/deep-agents/troubleshooting.md): Common issues and fixes for the Deep Agents SDK
  - [LangChain SDK (Python)](https://docs.openbox.ai/developer-guide/langchain.md): LangChain SDK reference and guide set
    - [Integration Walkthrough](https://docs.openbox.ai/developer-guide/langchain/integration-walkthrough.md): LangChain Python integration walkthrough
    - [Configuration](https://docs.openbox.ai/developer-guide/langchain/configuration.md): All LangChain SDK configuration options
    - [Error Handling](https://docs.openbox.ai/developer-guide/langchain/error-handling.md): Error codes and recovery patterns for the LangChain SDK
    - [Event Model](https://docs.openbox.ai/developer-guide/langchain/event-model.md): LangChain SDK event model and payload guidance
    - [Approvals and Guardrails](https://docs.openbox.ai/developer-guide/langchain/approvals-and-guardrails.md): LangChain SDK approvals and guardrails behavior
    - [Telemetry](https://docs.openbox.ai/developer-guide/langchain/telemetry.md): LangChain SDK telemetry capture and runtime behavior
    - [Troubleshooting](https://docs.openbox.ai/developer-guide/langchain/troubleshooting.md): Common issues and fixes for the LangChain SDK
  - [LangGraph SDK (Python)](https://docs.openbox.ai/developer-guide/langgraph.md): LangGraph Python SDK reference and architecture
    - [Configuration](https://docs.openbox.ai/developer-guide/langgraph/configuration.md): All LangGraph SDK configuration options
    - [Error Handling](https://docs.openbox.ai/developer-guide/langgraph/error-handling.md): Error codes and recovery patterns for the LangGraph SDK
    - [Integration Walkthrough](https://docs.openbox.ai/developer-guide/langgraph/integration-walkthrough.md): LangGraph-specific integration walkthrough
    - [Event Model](https://docs.openbox.ai/developer-guide/langgraph/event-model.md): LangGraph SDK event model and payload guidance
    - [Approvals and Guardrails](https://docs.openbox.ai/developer-guide/langgraph/approvals-and-guardrails.md): LangGraph SDK approvals and guardrails behavior
    - [Telemetry](https://docs.openbox.ai/developer-guide/langgraph/telemetry.md): LangGraph SDK telemetry capture and runtime behavior
    - [Troubleshooting](https://docs.openbox.ai/developer-guide/langgraph/troubleshooting.md): LangGraph SDK troubleshooting guide
  - [Mastra SDK (TypeScript)](https://docs.openbox.ai/developer-guide/mastra.md): Mastra SDK reference and guide set
    - [Configuration](https://docs.openbox.ai/developer-guide/mastra/configuration.md): Mastra SDK configuration reference
    - [Error Handling](https://docs.openbox.ai/developer-guide/mastra/error-handling.md): Mastra SDK error handling and runtime recovery patterns
    - [Integration Walkthrough](https://docs.openbox.ai/developer-guide/mastra/integration-walkthrough.md): Mastra-specific integration walkthrough
    - [Event Model](https://docs.openbox.ai/developer-guide/mastra/event-model.md): Mastra SDK event model and payload guidance
    - [Approvals and Guardrails](https://docs.openbox.ai/developer-guide/mastra/approvals-and-guardrails.md): Mastra SDK approvals and guardrails behavior
    - [Telemetry](https://docs.openbox.ai/developer-guide/mastra/telemetry.md): Mastra SDK telemetry capture and runtime behavior
    - [Troubleshooting](https://docs.openbox.ai/developer-guide/mastra/troubleshooting.md): Mastra SDK troubleshooting guide
  - [n8n Node Reference](https://docs.openbox.ai/developer-guide/n8n.md): n8n community node reference and guide set
    - [Integration Walkthrough](https://docs.openbox.ai/developer-guide/n8n/integration-walkthrough.md): n8n integration walkthrough
    - [Configuration](https://docs.openbox.ai/developer-guide/n8n/configuration.md): n8n node configuration options
    - [Error Handling](https://docs.openbox.ai/developer-guide/n8n/error-handling.md): Error handling and recovery patterns for the n8n node
    - [Event Model](https://docs.openbox.ai/developer-guide/n8n/event-model.md): n8n node event model and payload guidance
    - [Approvals and Guardrails](https://docs.openbox.ai/developer-guide/n8n/approvals-and-guardrails.md): n8n node approvals and guardrails behavior
    - [Telemetry](https://docs.openbox.ai/developer-guide/n8n/telemetry.md): n8n node telemetry capture and runtime behavior
    - [Troubleshooting](https://docs.openbox.ai/developer-guide/n8n/troubleshooting.md): Common issues and fixes for the n8n node
  - [OpenClaw Plugin](https://docs.openbox.ai/developer-guide/openclaw.md): OpenClaw governance plugin reference and overview (coming soon)
  - [Temporal Plugin (Python)](https://docs.openbox.ai/developer-guide/temporal-python/sdk-reference.md): Full plugin API reference
    - [Configuration](https://docs.openbox.ai/developer-guide/temporal-python/configuration.md): All plugin configuration options
    - [Error Handling](https://docs.openbox.ai/developer-guide/temporal-python/error-handling.md): Error codes and recovery patterns
    - [Temporal Integration Guide (Python)](https://docs.openbox.ai/developer-guide/temporal-python/integration-walkthrough.md): Python-specific Temporal integration guide
    - [Troubleshooting](https://docs.openbox.ai/developer-guide/temporal-python/troubleshooting.md): Common errors and fixes for the Temporal Python integration
      - [Concept](https://docs.openbox.ai/developer-guide/temporal-python/concept.md): Sandbox routing, providers, evidence, and fail-closed behavior
      - [Quick Start](https://docs.openbox.ai/developer-guide/temporal-python/quick-start.md): Fast path. Provision the native sandbox, then run a Temporal Worker whose activity is routed into it.
      - [Provisioning](https://docs.openbox.ai/developer-guide/temporal-python/provisioning.md): Provider selection, policy templates, release verification, and Worker environment
        - [Native Provider](https://docs.openbox.ai/developer-guide/temporal-python/native-provider.md): Native Seatbelt and bubblewrap provisioning, policy, egress, evidence, and limitations
        - [OpenShell Provider (VM)](https://docs.openbox.ai/developer-guide/temporal-python/openshell-provider.md): OpenShell microVM provisioning, prepared caches, zot registry, CA trust, and platforms
      - [Command Profiles](https://docs.openbox.ai/developer-guide/temporal-python/command-profiles.md): Registry reference. Argument types, result schemas, and admission rules for governed commands.
      - [Console Evidence](https://docs.openbox.ai/developer-guide/temporal-python/console-evidence.md): Evidence keys, result limits, and evidence limits for sandbox execution spans
    - [Extending the Demo Agent](https://docs.openbox.ai/developer-guide/temporal-python/customizing-the-demo.md): Customizing the demo for your use case
    - [Demo Architecture Reference](https://docs.openbox.ai/developer-guide/temporal-python/demo-architecture.md): How the demo agent is structured
  - [Working with llms.txt](https://docs.openbox.ai/developer-guide/llms-txt.md): LLM-optimized plain text documentation files

## Dashboard
- [Dashboard](https://docs.openbox.ai/dashboard.md): Trust overview and organization-wide monitoring
  - [Agents](https://docs.openbox.ai/dashboard/agents.md): Managing agents in the dashboard
    - [Registering Agents](https://docs.openbox.ai/dashboard/agents/registering-agents.md): How to register a new agent
    - [Agent Settings](https://docs.openbox.ai/dashboard/agents/agent-settings.md): Configure agent details and risk profile
  - [Inventory](https://docs.openbox.ai/dashboard/inventory.md): Org-wide discovery registry for agents, models, tools, and integrations, including unregistered ("shadow AI") callers
  - [Resource Catalog](https://docs.openbox.ai/dashboard/resource-catalog.md): Org-wide catalog of declared resources and agent role grants for the Agent IAM Gate
  - [Projects](https://docs.openbox.ai/dashboard/projects.md): Projects and agent lineage in the OpenBox dashboard
  - [Trust Overview](https://docs.openbox.ai/dashboard/trust-overview.md): Aggregate trust view across all agents
  - [Alerts](https://docs.openbox.ai/dashboard/alerts.md): Configuring and responding to alerts

## Administration
- [Administration](https://docs.openbox.ai/administration.md): Organization management and compliance overview
  - [Organization](https://docs.openbox.ai/administration/organization.md): Managing your OpenBox organization
    - [Teams](https://docs.openbox.ai/administration/organization/teams.md): Teams for grouping agents and scoping access
      - [Multi-Agent Sessions](https://docs.openbox.ai/administration/organization/teams/multi-agent-sessions.md): Per-team multi-agent run visualization
  - [Platform Operations](https://docs.openbox.ai/administration/platform-operations.md): Webhooks subscription API, platform-level API keys, and feature flags, distinct from per-agent credentials and notification integrations
  - [Identity Bridge](https://docs.openbox.ai/administration/identity-bridge.md): Opt-in CAEP-over-SSF signal emission on trust incidents, with Okta ITP as the first validated receiver
  - [Compliance & Audit](https://docs.openbox.ai/administration/compliance-and-audit.md): Compliance tooling and audit support
  - [Attestation & Cryptographic Proof](https://docs.openbox.ai/administration/attestation-and-cryptographic-proof.md): Tamper-evident cryptographic audit trails
  - [Audit Log](https://docs.openbox.ai/administration/organization-audit-log.md): Full org-level audit log reference