Harshit-J004/toolguard
Pytest-style reliability testing for AI agent tool chains. Catches hallucinated payloads, schema errors, and cascading failures before production.
Provides automated fuzzing and DAG-based execution tracing that identifies tool-chain vulnerabilities without live LLM calls—instead using type hints to generate deterministic failure scenarios (null propagation, type mismatches, cascading errors). Integrates natively with LangChain, CrewAI, Swarm, and AutoGen through context vars instrumentation, while a 6-layer security firewall adds human-in-the-loop approval gates for high-risk tool execution and recursive prompt-injection detection.
Stars
5
Forks
1
Language
Python
License
MIT
Category
Last pushed
Mar 17, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/agents/Harshit-J004/toolguard"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
valmi-io/value
⚡ "Value" - https://value.valmi.io . Valmi Value is Outcome-based billing and payments...
aptible/unpage
Unpage is the open source framework for building SRE agents with infrastructure context and...
Elijas/token-throttle
Multi-resource rate limiting for LLM APIs. Reserve tokens before you call, refund what you don't...
dipampaul17/AgentGuard
Real-time guardrail that shows token spend & kills runaway LLM/agent loops.
2001Haru/TokenWaster
The Most Useless EVER agent assistant in the Human History. Always trying to read everything in...