AGENT
NEMESIS

V 1.0.0Hackathon BuildEST 2026

A trust-auditing dashboard that reads real OpenTelemetry traces to catch loops, unverified claims, and broken promises before they burn your budget.

Enter Dashboard
Built onSigNozOpenTelemetry
DETECT LOOPS·VERIFY CLAIMS·AUDIT HANDOFFS·PREVENT HALLUCINATIONS·
DETECT LOOPS·VERIFY CLAIMS·AUDIT HANDOFFS·PREVENT HALLUCINATIONS·
DETECT LOOPS·VERIFY CLAIMS·AUDIT HANDOFFS·PREVENT HALLUCINATIONS·
DETECT LOOPS·VERIFY CLAIMS·AUDIT HANDOFFS·PREVENT HALLUCINATIONS·

WE EXPOSE
THE GAP

Core Capabilities

Infinite
Loops

Agents get stuck calling the same tool repeatedly without making progress. We detect cyclical tool usage and flag it before your API costs explode.

*
!

Unverified
Claims

An agent states something as fact without invoking a tool to check. We cross-reference final output with trace history to ensure cryptographic proof of work.

Broken
Promises

The agent claims to have executed an action but never dispatched the request. We trace intent vs execution to ensure reliability.

/

PIPELINE
JOURNEY

Step 01

Emit Traces

OTel GenAI conventions send deterministic spans to SigNoz continuously.

Step 02

Checker Worker

A headless Node process polls SigNoz, detecting logical gaps in the trace history.

Step 03

The Truth

Dashboard visualizes trust scores and flags, exposing what your agent actually did.

Start