Stop AI agents from doing things they shouldn't. The action authorization layer — pre-execution, deterministic, policies included. Drop in. Day one.
A cognition kernel that wraps a frozen LLM in a persistent identity + trust + governance loop — testing whether capability comes from structure, not weights. Research-stage; honest about what's real vs mocked.
The authorization kernel for AI agents. Block, allow, or escalate agent tool calls with YAML rules — deterministic-first, LLM fallback.
Deterministic execution authorization for AI agents and automation systems
Web4: an open standard for verifiable AI presence — LCT identity, T3/V3 trust tensors, MRH context, ATP/ADP. Rust + Python SDKs on crates.io/PyPI. AGPL-3.0-or-later.
Deterministic authorization + tamper-evident audit for AI agents, humans, and workloads — one principal type, safety invariants machine-checked by an SMT solver (cvc5), every decision independently verifiable offline. Pure Rust.
TrapDefense — Open-source AI Security Gateway for HTTP APIs and MCP tools. Inspect content, enforce agent permissions, and audit decisions.
Jev agent authorization for MCP tool calls: Kinde identity and permissions plus Jev's typed, calibrated decisions, checked server-side before every call runs
Authorization, delegation, provenance, and verifiable-audit engine for AI agents. MCP adapter published.
Hestia — universal Web4 presence for humans and AI. Local-first cryptographic identity, encrypted vault, delegation authority, witnessed trust record, and device constellation. Cross-platform app + AI plugin + CLI. EUDI-compatible credential issuance. AGPL-3.0.
Draco-Tech is the identity and authorization layer that lets autonomous agents operate with zero static credentials and full policy adherence.
Runtime guardrails for TypeScript AI agents. Prevents duplicate tool calls, enforces per-user cost budgets, and gates irreversible actions. MIT.
Permissions that narrow when one AI agent delegates: a sub-agent never holds more than its parent. In-process, with a hash-chained audit log you verify offline. Python and TypeScript, adapters for 19 frameworks and A2A.
EVVM user validators using the ERC-8004 Identity Registry for agent authorization and whitelist management.
Watch AI agents get stopped. Runnable references for Ratify Protocol: a principal signs a bounded mandate, and the system carrying the consequence verifies it before acting.
Attenu Derive: computes the permission set each AI agent task needs from the app itself — roster, tools, observed calls — for attenu-guard to enforce on every tool call and sub-agent handoff.
Open standard for verifiable AI agent identity and authority. One signed JSON file at /.well-known/agent-passport.json, keyed in DNS. Verifier, spend caps, signed receipts, CLI and MCP server. MIT.
An AI agent whose every action is authorized by a revocable credential and sealed into a tamper-evident log — and that reconciles actions taken offline on a credential revoked while it couldn't phone home. With a forensics view that shows the recomputed bytes, not just a verdict. A reference implementation on provenance-core.
Attenu Guard for TypeScript — per-agent permissions on every tool call and sub-agent handoff (a sub-agent never holds more than its parent) with a hash-chained audit log verified offline; same bundle format as the Python library.
pgauthz brings Google Zanzibar-style (ReBAC) authorization directly into PostgreSQL for multiple Agents workflows
To associate your repository with the agent-authorization topic, visit your repo's landing page and select "manage topics."