Trust scoring for GitHub PR authors based on contribution history.

🧠 Local-first knowledge system for LLM agents — sqlite-vec + ONNX embeddings, no cloud/Docker/PyTorch dependency

NEXO runtime core for NEXO Desktop: local memory, automation, MCP tools and update-managed runtime.

An operating model for autonomous agent teams. Identity, task contracts, failure memory, trust scoring and audit for teams that want to operate agents as an accountable workforce.

Community-driven behavioral reliability benchmark for LLMs. 231 probes across 19 modules, deterministic scoring, perplexity correlation, layer sensitivity mapping, quant method capture, hardware-stratified community rankings. Every test contributes to the community dataset.

Memory defense for AI agents — stops MINJA, AgentPoison, and MemoryGraft attacks. Zero dependencies.

Cryptographic verification for AI agent actions — ECDSA-secp256k1 + RFC 6979 signed Action Receipts (v0.1), multi-dimensional Trust Vector, capability tokens (JWT-shaped + Biscuit-style attenuation), UETA §10(b) undo. 29 MCP tools. A2A v1.0 + ERC-8004 format compatible. Receipt-batch Merkle roots anchored on Base mainnet. Starting with code.

Bayesian change-trust layer. Autonomously scores every AI code edit in real-time, instantly flagging the exact files that demand human review.

Official SDKs and MCP server for VeriSwarm — trust infrastructure for AI agents

MCP package manager with built-in trust scoring. Search, install, and audit MCP servers across Claude Desktop, Cursor, VS Code, and Windsurf.

Persistent, trust-scored project memory for AI coding agents — MCP server + CLI, backed by PostgreSQL + pgvector.

TRUCE Agent Trust Framework (TATF) — Open standard for autonomous agent trust scoring. Apache 2.0.

Exact Bayesian trust and reputation scoring for tools, MCP servers, skills, and agents. Zero-dependency, TypeScript-first: closed-form Beta-Bernoulli posteriors, calibrated credible intervals, Thompson-sampling routing, time-decay, and a tamper-evident audit trail.

Real-time blockchain fraud detection using autonomous multi-agent AI swarms. Detect fraud in <2s with 150x faster vector search (HNSW), 84.8% accuracy, and 73% cost reduction. Features ERC-1155 trust scores, MCP integration, and reflexion learning.

Structured fact memory MCP server — SQLite + FTS5, trust scoring, entity graph, auto-observe, bilingual retrieval for AI coding assistants

AI-native version control system that extends Git for autonomous agent workflows. Tracks decisions, not just diffs. Tournament-style parallel execution, trust-based reputation scoring, and full chain-of-thought lineage.

Explainable AI-Powered URL Trust Intelligence Engine — 15+ parallel analysis engines, hybrid 70/30 rule+AI scoring, brand impersonation detection, behavioral analysis, and full-transparency evidence breakdown. Self-hosted, open-source.

Multi-agent orchestration with self-evolution, cognitive memory, and governance.

Open-source Rust runtime for autonomous agents — 13 modules, 76 crates, 2625 tests. Event-sourced persistence (Lago), homeostatic regulation (Autonomic), x402 payments (Haima), cryptographic identity (Anima), distributed networking (Spaces), metacognitive evaluation (Nous), OpenTelemetry observability (Vigil).

TAMR+: Trust-Aware Multi-Signal Document Retrieval for Regulatory AI. TRACE scoring, gap attribution, 250-question benchmarks, HashGNN embeddings. EU AI Act compliant. Patent EP26162901.8.