cap|fence
WhyPlatformDocsPatternsBlog
v0.1 · private betaGitHub
Documentation

Docs

CapFenceMigration Guide: Gate to ActionRuntime

Reference

ActionRuntime & ActionEvent API ReferenceAudit Log FormatCLI ReferenceConfiguration ReferenceFlowTracer API ReferencePolicy Schema Reference

Guides

Air-Gapped DeploymentsCI/CD EnforcementObserve Mode RolloutProtect Payment AgentsProtect Shell ToolsReplay an IncidentRequire Human ApprovalSecure MCP Servers

Architecture

Approval LifecycleAudit Chain DesignEnforcement FlowFailure BehaviorInternal ComponentsMCP Interception ModelPolicy Evaluation ModelReplay Engine ModelThreat Model

Examples

Approval Workflows ExampleDatabase Write GatingDemo CastDemo WalkthroughFintech Agent ExampleMCP Governance ExampleReplay Demo

Concepts

Audit ChainFail-Closed EnforcementPolicy ModelReplayabilityRuntime AuthorizationTrust Propagation

Integrations

AutoGen IntegrationCompatibility MatrixCrewAI IntegrationCustom Framework IntegrationLangChain IntegrationLangGraph IntegrationLlamaIndex IntegrationMCP IntegrationOpenAI Agents SDK IntegrationPydanticAI Integration

Compliance

EU AI Act EvidenceGovernance ReportingOWASP Agentic Top 10

Getting Started

InstallationQuickstartYour First Blocked ActionYour First Policy

Recipes

Protect Payments With ThresholdsProtect Production DB WritesProtect Shell ToolsRecipesRequire Approval For SaaS Admin ChangesSecure MCP Filesystem Access

Recipes

Copy-paste recipes for common protection patterns. Each recipe includes policy YAML and a minimal integration snippet.

  • Protect shell tools
  • Protect production DB writes
  • Protect payments with thresholds
  • Secure MCP filesystem access
  • Require approval for SaaS admin changes
cap|fence

Execution authorization infrastructure for autonomous systems.

DocsPatternsBlogGitHub