Claim Matrix Console (K3)

An agent-swarm system for mapping contradictory data feeds into context graphs.

Claim Matrix Console (K3) media is blocked

Allow external media to connect to the provider and play this content.

Description

K3 is an AI agent system for 'Context Graph Engineering' that preserves source contradictions rather than averaging them into a single summary. It uses an agent-swarm architecture to map data feeds into nodes and claims, using 'forked nodes' to retain multiple values and tagging relationships as corroboration, scope mismatch, stale data, or real spread.

Descriptions, tags, and model credits may be AI-generated or inferred from public sources and can be incomplete or wrong. Learn more.

Details

Why this is interesting: It addresses the critical 'averaging' problem in LLM summarization by architecting a graph that explicitly preserves…
Built with: Kimi K3
Tags: #agents #knowledge-graph #data-verification #multi-agent

Discovered: X
First posted: X · Aug 17, 2026
Listed: Added 9d ago
Build: Moderate
Models: Strong

Prompt

TASK: map EV battery supply risk. one agent per source.
SOURCES: Yahoo Finance, World Bank, IMF, company filings, live stocks.

NODE: one company. match on ticker.
FIELDS: see fields.md. record unit, scope, period with every value.

CONNECT: attach every claim to the entity it describes.
         link two claims that match as corroboration.

CONFLICT: when two claims about one field disagree, keep both.
          tag the edge with the reason: value, date, or scope.
          do not average. do not pick a winner.

OUTPUT: node count, claim count, conflict count, then the summary.

Source

All projects

Agents / Skills