Skip to content

SponsioLabs/Sponsio](https:

AI Agents & Assistants

Runtime enforcement tool that blocks unsafe actions of AI agents with deterministic code contracts before they execute

Python Latest v0.1.1 · 12d ago Security brief →

Features

  • Enforces agent actions against deterministic, pure‑code contracts at runtime
  • Zero LLM cost in the hot path; average latency ~0.14 ms per check (5k–60k× faster than LLM judges)
  • Works with major frameworks: LangChain, Claude Agent, OpenAI Agents, Google ADK, CrewAI, Vercel AI, MCP and custom loops
  • Supports both Python and TypeScript projects via CLI wizard or one‑shot prompts
  • Provides audit trails, pass/block/warn/escalate decisions and detailed benchmark results

Recent releases

View all 2 releases →
Upgrade now
v0.1.1 Bugfix
Breaking upgrade

Fixes pyyaml crash

Upgrade now
v0.1.0 Bug fix
Breaking upgrade Auth RBAC

Restored missing modules

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

About

Stars
470
Forks
28
Languages
Python TypeScript JavaScript
Downloads/week
3 ↓95%
NPM Maintainers
1 Single npm maintainer
Contributors
4
TypeScript
Types included ✓

Install & Platforms

Install via
pip npm

Community & Support

Beta — feedback welcome: [email protected]