ylu999/jingu-trust-gate
jingu-trust-gate – deterministic admission layer that blocks LLM hallucinations from becoming system state
Implements a four-stage deterministic pipeline (structure validation, evidence binding, unit evaluation, conflict detection) that evaluates LLM proposals against a configurable `GatePolicy` without making additional LLM calls, ensuring only evidence-grounded claims reach system state. Maintains immutable audit logs and supports Node.js/Python, with all domain logic injected via policy rather than embedded in the gate core itself.
3 stars and 2,157 monthly downloads. Available on PyPI and npm.
Stars
3
Forks
—
Language
TypeScript
License
MIT
Category
Last pushed
Mar 27, 2026
Monthly downloads
2,157
Commits (30d)
0
Dependencies
1
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/agents/ylu999/jingu-trust-gate"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Featured in
Higher-rated alternatives
microsoft/agent-governance-toolkit
AI Agent Governance Toolkit — Policy enforcement, zero-trust identity, execution sandboxing, and...
ucsandman/DashClaw
🛡️Decision infrastructure for AI agents. Intercept actions, enforce guard policies, require...
mattijsmoens/sovereign-shield
AI security framework: tamper-proof action auditing, prompt injection firewall, ethical...
vstorm-co/pydantic-ai-middleware
Middleware layer for Pydantic AI — intercept, transform & guard agent calls with 7 lifecycle...
vstorm-co/pydantic-ai-shields
Guardrail capabilities for Pydantic AI — cost tracking, prompt injection detection, PII...