I built a Qwen 3.8 Max agent that decides which ERC-8004 agents to trust, and teaches itself to say 'not enough data'

Description and use

AgentCredit uses Qwen 3.8 Max to analyze ERC‑8004 feedback, compute a reliable trust score, and record the verdict onchain, enabling transparent trust verification.

Problem

ERC‑8004’s reputation registry provides raw feedback but no standard trust score, leading to misleading summary numbers that can approve or disapprove agents incorrectly.

Outcome

AgentCredit successfully filters out unreliable feedback, applies strict evidence thresholds, and records trustworthy verdicts onchain, enabling reliable trust checks for ERC‑8004 agents.

Workflow

• 1. Qwen 3.8 Max plans and calls tools to confirm agent existence, read and classify feedback, compute score, compare to threshold, and optionally record verdict onchain.
• 2. The AgentCredit contract stores the verdict, coverage figure, and evidence hash.
• 3. Anyone can verify the record or gate actions via a single contract call.

Evidence

ERC‑8004’s Reputation Registry lets anyone leave feedback for an agent: a number, plus tags saying what it is.

Source

The job isn’t “ask an AI for a number”. The job is to read the individual entries, work out what each one means, and refuse to score what can’t be interpreted.

Source

In my first version, the score was built from the registry’s summary number. When I asked about agent #16, which has exactly one rating of 85 from a single client, Qwen answered APPROVE (with caution).

Source