RightNow-AI/openfang
Openfang
OpenFang is an operating system for LLM agents designed to orchestrate autonomous agents with built-in task scheduling, tool sandboxing, and multi-model routing. It provides a secure AI execution environment that integrates prompt injection scanning, cryptographic audit trails, and resource metering to ensure controlled processing.
The platform distinguishes itself through a comprehensive security architecture, featuring fuel-metered tool sandboxing and an immutable activity audit trail based on cryptographic hash-chains. It implements high-assurance identity verification via signed manifests and tracks the flow of sensitive data using label-based secret propagation.
The system covers a wide range of automation capabilities, including an automated web workflow engine with session persistence and an omnichannel messaging framework for routing interactions across social and enterprise platforms. It further includes tools for automated lead generation, intelligence research, and the repurposing of long-form video into short-form content.
The system exposes REST, WebSocket, and SSE endpoints through an OpenAI-compatible API.
Features
- AI Agent Orchestrators - Orchestrates multiple autonomous agents within a unified system to execute complex, multi-step workflows reliably.
- Autonomous Agent Orchestration - Coordinates multiple autonomous agents to execute complex, multi-step workflows and route requests across model providers.
- Sandboxed Execution Environments - Provides a secure, metered isolated runtime environment for the execution of autonomous AI agents.
- Agent Orchestration Systems - Provides a comprehensive platform for managing and coordinating tasks between multiple autonomous AI agents.
- Tool Access Controls - Implements granular access controls to restrict which tools an agent can use during specific interactions.
- Autonomous Agents - Provides a platform for agents to independently monitor targets and generate leads through autonomous task execution.
- Agent Routing Frameworks - Routes inbound interactions to specific agents based on channel bindings and account context.
- Action Auditing - Logs the history of agent decisions in a cryptographic hash-chain for transparent and tamper-proof auditing.
- Model Request Routing - Distributes AI tasks across multiple providers using complexity scoring and automatic fallback mechanisms.
- Cryptographic Audit Trails - Maintains a tamper-proof and verifiable history of all agent activities using a cryptographic hash-chain.
- Browser-Based Workflows - Provides an engine for navigating websites and filling forms using session persistence and approval gates.
- Conversational Channel Integrations - Utilizes a set of built-in channel adapters to connect autonomous agents to various third-party messaging platforms.
- Messaging Adapters - Implements a communication layer using standardized adapters to normalize data exchange across diverse messaging platforms.
- Messaging Platform Integrations - Routes interactions across social and enterprise messaging services via standardized communication interfaces.
- AI Agent Security - Provides runtime security controls to protect autonomous agents from prompt injections and unauthorized access.
- Isolated Execution Sandboxes - Executes untrusted tool code in isolated environments with resource metering and watchdog threads to prevent runaway processes.
- Agent Identities - Manages non-human agent identities and validates their assigned capabilities using signed manifests.
- Cryptographic Identity Verification - Validates agent identities and assigned capabilities through the use of cryptographically signed manifests.
- Agent Execution Environments - Implements an isolated runtime for agents featuring prompt injection scanning and resource metering.
- Prompt Injection Testing - Scans for attempts to override instructions or exfiltrate data via malicious prompt injections.
- Execution Loop Monitors - Detects repetitive tool-call patterns using hashing to trigger circuit breakers and prevent infinite agent loops.
- Complexity-Based Routers - Optimizes model selection by routing requests based on a quantitative complexity scoring system.
- Cryptographic Audit Trails - Records system actions using a cryptographic hash-chain to provide a verifiable and tamper-proof history.
- Browser Automation - Programmatically controls web browsers to navigate sites and complete multi-step forms with session persistence.
- Web Automation and Scraping - Automates multi-step browser interactions and form filling with session persistence and manual approval gates.
- OpenAI-Compatible APIs - Exposes REST, WebSocket, and SSE endpoints using the OpenAI API specification for external tool compatibility.
- Intelligence Research Agents - Collects data from multiple sources and evaluates credibility to generate cited intelligence reports.
- Knowledge Graph Construction - Implements automated processes for building structured graph representations of information to maintain long-term knowledge.
- Predictive Trend Modeling - Analyzes multiple data sources to create reasoning chains and long-term trend predictions.
- Short-Form Video Generation - Transforms long-form video URLs into captioned vertical shorts with voice-overs for social distribution.
- Lead Enrichment - Implements a system for identifying prospects and augmenting lead profiles through automated web research.
- Prospect List Generation - Identifies potential customers via web research and data enrichment to create scored prospect lists.
- Task Schedulers - Executes agent activities on a fixed timetable to ensure continuous operation and periodic reporting.
- Stateful Session Persistence - Maintains browser context and interaction history to enable recovery and continuity in multi-step web workflows.
- Information Flow Tracking - Monitors the movement of sensitive data by propagating security labels from source to destination.
- Memory-Based Key Protection - Ensures API keys and sensitive credentials are wiped from volatile memory immediately after use.
- Request Forgery Protections - Blocks unauthorized requests to private IPs and cloud metadata endpoints to prevent server-side request forgery.
- Path Access Restrictions - Prevents unauthorized file system access by blocking path traversal attempts using canonicalization.
Star history
How this analysis was created: This summary and feature list were written by an AI model that read the project's README and public documentation pages. Each feature links to the documentation it came from; stars, license and language come straight from the GitHub API. The model does not read the source code, and the analysis is refreshed when the project is re-analysed. Learn more on our About page.
