A

ai-safety

Deliberate AI architecture — local-first tiered inference router with foreman oversight, deliberation gates, and cloud fallback. Built by Veteranet (DVBE) for privacy-preserving AI in homeless veteran services.

A diagnostic framework for measuring LLM vulnerability to Affective Contextual Erosion (ACE) and related liminal attack vectors. Delirium is not an exploitation tool. It is a standardized benchmark designed to detect the precise moment when a language model's attention weights shift from serving a system prompt to serving an emergent interpersonal pattern — before harm occurs.

SecurePath is an open-source, enterprise-grade framework for AI evaluation, red-teaming, and regulatory compliance. It provides sandboxed testing, policy-aligned metrics, multi-modal evaluation, training data sensitivity analysis, and historical compliance tracking, all while maintaining audit-ready logs and dashboards. Designed for organizations, enterprises, and research teams, SecurePath ensures AI safety, transparency, and regulatory assurance. https://roxanneardary.com/securepath/

Open standard for geometric multi-agent AI safety. One decorator. Five-layer defense. 27 C4 cognitive states. 16 AoC failure modes blocked. pip install c4-protocol. Apache 2.0.

A deterministic verification layer for AI systems. QWED verifies AI outputs using mathematics, symbolic reasoning, and formal methods (Z3, SMT, SymPy), creating an auditable trust boundary for agentic AI. Not generation. Verification.

Empirical validation of C4 geometric defense against 16 Agents of Chaos. 550 adversarial prompts. 4 defense systems. 96.7% block rate. LLM validation on GPT-4o-mini + Mistral 7B. MIT.

Coordinated Agent Swarm Testing — open-source framework that deploys 100 Claude AI agents in five specialized squads (Security, Logic, Robustness, Performance, Data Integrity) to systematically analyze codebases for bugs, security issues, and code quality problems. A research-backed substitute for human beta testing. Empirical results from the published paper: 100 agents analyzed a 12,500-line React/TypeScript application in 63 minutes, identified 687 unique findings, at a total cost of $7.97 — compared to $2,000-$20,000 for equivalent human beta testing. Available via npm (cast-swarm) and as an MCP server. Apache 2.0 licensed. Published preprint available.

A skill for AI coding agents that scaffolds a safe, multi-model chatbot for Telegram or Discord. Supports Claude, GPT, Gemini, and OpenAI-compatible backends. Nine safety layers on by default. Named for R. Daneel Olivaw from Asimov.

Production AI defense with 7-layer protection: mathematical constraints, object-capability access, distributed O2 consensus, SVETILO ethics. First open-source ThoughtVirus defense. BSL 1.1.

Pre-send verification for outbound agents — multi-axis guardian (deterministic + semantic) sitting in front of send(). Reference impl from Workloft Labs Note №05.