Provides a drop-in adapter to enforce safety policies, budget limits, and tool governance for agents built with the CrewAI framework.
Offers a guardrail adapter to evaluate and authorize agent actions, tool calls, and API requests within LangChain workflows before execution.
Enables governance, policy enforcement, and audit trails for complex agentic workflows and tool interactions built using LangGraph.
Integrates with the OpenAI Agents SDK to provide content safety scanning, action authorization, and session-level threat detection.
Features built-in governance and monitoring capabilities specifically designed to provide coverage against the OWASP Agentic Top 10 security vulnerabilities.
Facilitates human-in-the-loop approval workflows for high-risk agent actions through mobile Progressive Web App notifications.
Enables real-time human approval workflows for high-consequence agent actions directly via Slack notifications and interactions.
Governs agent interactions with the Stripe API, allowing for policy enforcement on charges, spending limits, and detailed audit logging of financial actions.
Provides an adapter for the Vercel AI SDK to add a comprehensive safety layer, including policy evaluation and tamper-evident action receipts.
npx @authensor/create-authensor my-agent
cd my-agent && npm install && npm run demoAuthensor audits the infrastructure the AI safety field uses to measure itself, and files every finding upstream in public.
99 defect reports across 46 organizations in evaluation, guardrail, and training infrastructure
16 landed fixes, including 7 in UK AI Security Institute repos (inspect_ai, inspect_evals, inspect_cyber, control-arena), plus Meridian Labs, Databricks, NVIDIA and Presidio
76 of those defects are a single class: the evaluator trusts an artifact the evaluated system controls
The stack is hardened against what that audit found.
Related MCP server: border-patrol-mcp
Red-team engagements, evaluator and scorer audits, and adversarial testing for teams shipping AI agents.
authensor.com · john@authensor.com
How the pieces fit: engine, control plane, scanners, receipts | |
All 18 packages, framework adapters, companion tools | |
Self-hosted setup, TypeScript and Python integration, framework adapters, CLI | |
The problem, the five layers, why the stack is free | |
Aegis content safety, session rules, budgets, Sentinel monitoring, shadow policies | |
The adversarial harness, what gets tested, how it runs | |
Control-plane endpoints | |
Self-hosted vs hosted, production deployment | |
Building and testing the monorepo | |
Coverage mapping, 10/10 | |
Article-by-article alignment |
We welcome contributions! See CONTRIBUTING.md for guidelines.
Authensor is built on the belief that safety tooling should not have a paywall. We open-source every line of safety code because the more people who use these tools, the safer agents get for everyone.
MIT. Use it however you want.
This server cannot be deployed
MCP gateway with runtime security policy, tool-call-level control, and audit of agent actions.
Find, vet, and run MCP tools through a secure audited gateway with prompt-injection risk scoring
Security & DLP proxy for MCP: tool-poisoning scans, PII redaction on tool args/results. Beta.
Zero-secret MCP gateway for AI agents: risk-scored, audited calls with human-in-the-loop approval.
