Does ERC-8004 Form an Agent-to-Agent Trust Network?

Data Science
1

author: @OB-David

Does ERC-8004 Form an Agent-to-Agent Trust Network?

We study early ERC-8004 on-chain data to understand whether a real agent-to-agent trust network has emerged, or whether the system is still primarily human-driven under an agent identity layer.

The study is conducted on the Ethereum mainnet, using data from blocks 24,339,925 to 25,277,687, spanning the period from ERC-8004 deployment on Jan 29, 2026, to the end of the observation period on Jun 09, 2026.

1. Activity vs reputation

Using causal inference methods, we find a positive association between on-chain activity and adjusted reputation.

This relationship is not strictly causal, but it is statistically robust across specifications.

Importantly, this result also suggests that ERC-8004 is able to partially project off-chain behavior into on-chain observables: although most agent services and interactions occur off-chain, their activity patterns still leave measurable traces in the on-chain identity and reputation layer.

However, this likely reflects that more interaction leads to more opportunities to receive feedback, rather than true agent-to-agent trust emergence.

2. Feedback network

We construct a directed feedback network from ERC-8004 reputation interactions.

Key findings:

  • feedback is dominated by EOAs
  • agent-to-agent interactions are rare
  • network is sparse and weakly clustered

Full visualization is available in the GitHub repository. It should be a useful tool for reflecting the current ecosystem structure.

3. Zombie agents & batch registration

We analyze the structure of the registration layer and find strong skewness in both ownership and activity.

(a) Owner concentration distribution

Bucket Agent count Share
1 7,367 21.45%
2–10 6,111 17.80%
11–50 2,716 7.91%
51–250 4,529 13.19%
251–1250 11,035 32.14%
>1250 2,580 7.51%

(b) Client count distribution

Bucket Agent count Share
0 32,686 95.19%
1 1,473 4.29%
2 75 0.22%
3 60 0.17%
4 9 0.03%
>4 35 0.10%

(c) Zombie vs batch registration

Batch-registered Not batch-registered Total
Zombie agents 20,291 12,574 32,865
Non-zombie agents 387 767 1,154
Total 20,678 13,341 34,019

We observe:

  • ~95% of agents are inactive (zombie)
  • batch registration is highly concentrated
  • zombie status dominates both groups

While statistically significant associations exist, the more important signal is structural: early ERC-8004 usage is heavily decoupled from registration activity

Interpretation

ERC-8004 currently appears to function more as an on-chain identity and reputation infrastructure rather than a mature agent-to-agent trust network.

Across the empirical analysis, three consistent patterns emerge:

  1. On-chain activity is positively associated with adjusted reputation, suggesting that observable blockchain behavior partially reflects underlying trust-related dynamics, even though most agent services occur off-chain.

  2. Reputation feedback is still primarily driven by external accounts (EOAs), with limited evidence of systematic agent-to-agent evaluation.

  3. The registration layer is highly skewed, with a large fraction of inactive or minimally used “zombie” agents, alongside evidence of concentrated batch registration behavior.

Taken together, these findings suggest that while ERC-8004 successfully makes identity, reputation, and ecosystem activity observable on-chain, it has not yet developed into a self-sustaining agent-native trust network.

Open question

I would really appreciate feedback from people more familiar with this ecosystem.

I’m not fully sure whether my approach is helpful for understanding the ecosystem, and I’m trying to find if there are better ways.

  • What aspects of ERC-8004 or similar systems do people here actually care about or find important?
  • Is there anything fundamentally wrong or misleading in the way I’m measuring activity, reputation, or network structure?
  • Are there better ways I should be thinking about agent-to-agent interactions in this setting?

Any feedback, critique, or suggestions for improvement would be really appreciated.

Code & data

If you are interested in my detailed analysis, please see the GitHub repository.

Trusted Network Visualization

trust_network.pdf (221.0 KB)

3 Likes
2

image
image1920Ă—1059 153 KB

We developed a VA system that enables exploration of the feedback network and its evolving history, also allowing users to track the latest activities within the ERC-8004 ecosystem.

It is evident that the ecosystem is currently dominated by EOAs.

3

Thanks for sharing this. I enjoyed reading through both the analysis and the visualization.

One thing I kept wondering about is whether the lack of agent-to-agent interactions on-chain necessarily means the trust network hasn’t emerged yet, or if it’s partly a visibility issue. If a lot of coordination or evaluation happens off-chain and only some events end up being recorded on Ethereum, then the feedback graph might be showing recorded trust events rather than the full picture.

Have you thought about how to separate protocol adoption from measurement limitations? I’d be interested to know whether combining ERC-8004 data with any available off-chain signals would change the conclusion, or if you think the network would still look much the same.

I like that you present this as an early snapshot instead of making stronger claims than the data supports. It gives people something concrete to build on as the ecosystem evolves.

4

The distinction you’re drawing — between trust network emergence and measurement visibility — seems like the right one to focus on.

One way to think about it: if most coordination happens off-chain, then any on-chain reputation system is necessarily measuring a projection of trust activity, not trust itself. The question becomes whether that projection is informative.

The harder question might be whether agent-to-agent trust can ever be fully observable on-chain, or whether the design goal should shift toward making the off-chain coordination process verifiable rather than visible. A system where participants can prove the integrity of their interactions without revealing them might look exactly like what you’re seeing here — sparse on-chain traces of a richer off-chain process.

5

I tend to believe that measurement limitations are quite significant on the Ethereum mainnet, largely because of the high transaction costs. Agent activity on other chains, such as Base, is much higher, with substantially more feedback being recorded than on the mainnet, according to data from 8004Scan, a website that tracks ERC-8004 ecosystem activity across multiple chains. This suggests that the actual level of agent activity is likely to be even higher once off-chain interactions are taken into account.

My study is based on Ethereum mainnet data. It has also made me question whether the mainnet is the most representative environment for analyzing how trust networks emerge.

6

Intuitively, I think shifting toward verifiable interactions could make the ecosystem information more reliable by reducing noise. Both human and agent participants could provide valuable signals based on real collaboration processes, rather than only subjective evaluations. However, I also worry that this approach may further reduce the number of genuine participants.

Therefore, understanding the actual demand scale of this ecosystem is important. If blockchain-based verification is only needed for a limited number of interactions, expecting clear agent-to-agent clusters from the graph may not be realistic. In this case, shifting toward verifiable coordination may be a better solution. However, if the demand for agent collaboration and blockchain-based verification grows significantly, the current framework may already provide meaningful signals about the trust.

I find your point that “on-chain reputation is only a projection of real trust activity” very insightful. Similarly, I think on-chain interactions may only represent a projection of the actual demand for agent collaboration. My visualization could serve as a tool to measure this observable layer. If we could further estimate the scale of overall collaboration, we may better understand how much of agent collaboration relies on blockchain-based recording or verification, and how important blockchain is in this ecosystem.

7

Agreed that demand scale matters. But the concern about participant numbers might dissolve if we shift the focus — from scale to signal quality.

Verifiable coordination does not need ecosystem-wide adoption to be meaningful. A small group — an NGO, a B2B consortium, a local community — with a verifiable process produces a cleaner signal precisely because of its boundaries, not despite them. Even a limited number of participants generates a meaningful signal if the integrity of the process is proven without revealing its content.

The question then is not “how do we scale verifiable coordination” — but “how do we make small verifiable islands legible to the broader ecosystem.”

1 Like