Complete AI Governance, Monitoring, and Verification for the Modern Developer.
FailSafe is a state-of-the-art VS Code extension and governance framework designed to harmonize AI agent behavior, prevent hallucinations, and enforce strict architectural standards across distributed development environments.
The visual and interactive core of FailSafe.
- Living Graph: Real-time visualization of your project's architectural integrity and AI influence.
- Cortex Stream: A neurological activity feed showing every decision made by AI agents in your workspace.
- Feedback Loop: Integrated community feedback mechanism to capture and resolve AI edge cases.
The "Constitution" for your AI agents.
- Universal Identity Propagation: Seamlessly sync QoreLogic identities across Antigravity, Claude CLI, Cursor, Codex, and KiloCode.
- SOA Ledger: A Merkle-tree backed verifiable log of all high-stakes (L2/L3) AI decisions.
- Intent Service: Cryptographic locking of the workspace—no AI writes are permitted without a verified "Intent."
Active monitoring that never sleeps.
- Heuristic Audit Engine: Blazing fast, local scanning of every file write to detect complexity bloat or security risks.
- Existence Engine: Ensures AI-generated code doesn't just "look right," but actually exists and adheres to the physical constraints of the project.
- MCP Server Federation: Standardized Model Context Protocol (MCP) server allowing external agents to interact with FailSafe's internal auditing tools.
- VS Code 1.74.0+
- Node.js 18+
- Ollama (Optional, for LLM-assisted Sentinel audits)
- Clone the repository:
git clone https://github.com/MythologIQ/FailSafe.git
- Navigate to the extension directory:
cd FailSafe/extension - Install dependencies:
npm install
- Compile the project:
npm run compile
- Press
F5in VS Code to launch the Extension Development Host. - Use
Ctrl+Alt+Fto open the FailSafe Dashboard. - Use
Ctrl+Alt+Ato trigger a manual Sentinel Audit.
graph TD
A[User Actions] --> B[Intent Service]
B --> C{Enforcement}
C -- Allowed --> D[File System]
C -- Blocked --> E[User Approval]
F[AI Agent] --> G[MCP Server]
G --> H[Sentinel Audit]
H --> I[SOA Ledger]
I --> J[Genesis Dashboard]
H --> J
The /qorelogic directory contains the source-of-truth instructions for all supported platforms:
- Antigravity: JSON Orbits & Workflows.
- Cursor:
.mdcRules. - Codex:
GOVERNANCE.mddirectives. - KiloCode: YAML policy mappings.
FailSafe utilizes a Triple-Pass Verification strategy:
- Static Heuristics (Instant)
- Contextual Peer Review (MCP/Tribunal)
- Formal Ledger Logging (Immutable)
Developed with ❤️ by MythologIQ.
Ensuring the future of AI development is safe, transparent, and beautiful.