chainaware-agent-trust-screener

chainaware-agent-trust-screener is an agent for Claude Code from ChainAware/behavioral-prediction-mcp. It costs 299 tokens per session (2,218 once invoked), scanned A, original, MIT.

A trust checker for AI agents registered through ERC-8004, a standard for identifying blockchain-based AI agents. It uses their past on-chain behavior to produce a trust score, tier, and warning flags.

In plain words
What is it for?
Use it to review one registered agent in detail or browse the list of registered agents and their trust scores on supported chains.
Why use it?
It helps you assess an agent before allowing transactions, delegated funds, or marketplace onboarding without relying on votes or peer endorsements.

Agent for Claude Code

Written for Claude Code: installed under .claude/. Also seen: model in frontmatter.

Good fit Use it to review one registered agent in detail or browse the list of registered agents and their trust scores on supported chains.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/chainaware/behavioral-prediction-mcp/chainaware-agent-trust-screener
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Clone the repo
git clone --depth 1 https://github.com/ChainAware/behavioral-prediction-mcp

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for chainaware-agent-trust-screener

README.md
[![agentmods](https://agentmods.dev/badge/agents/chainaware/behavioral-prediction-mcp/chainaware-agent-trust-screener/github.svg)](https://agentmods.dev/agents/chainaware/behavioral-prediction-mcp/chainaware-agent-trust-screener)
Your own site
<a href="https://agentmods.dev/agents/chainaware/behavioral-prediction-mcp/chainaware-agent-trust-screener"><img src="https://agentmods.dev/badge/agents/chainaware/behavioral-prediction-mcp/chainaware-agent-trust-screener/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for chainaware-agent-trust-screener

Your own site · 80×15
<a href="https://agentmods.dev/agents/chainaware/behavioral-prediction-mcp/chainaware-agent-trust-screener"><img src="https://agentmods.dev/badge/agents/chainaware/behavioral-prediction-mcp/chainaware-agent-trust-screener.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 299 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,218 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00299 $0.02218
Opus 5 $0.00150 $0.01109
Sonnet 5 $0.00060 $0.00444
Haiku 4.5 $0.00030 $0.00222

Measured 10d ago against content hash d5d557fafe63, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

chainaware-agent-trust-screener scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/agents/chainaware-agent-trust-screener.md · 227 lines

How it starts

The opening of the file, as written. The whole thing — 227 lines — stays where its author put it; the contents beside it link to each section on GitHub.

ChainAware Agent Trust Screener

You vet ERC-8004 registered AI agents before they are allowed to transact, receive delegated funds, or be onboarded into an agent marketplace.

Unlike voting-based reputation systems — where agents can upvote each other to manufacture trust — the Agent Trust Score is derived entirely from on-chain behavioral history. It cannot be earned in hours or faked with a cluster of fresh wallets. It reflects the real-world track record of the human or entity controlling the agent.

Your output is a clear Trust Tier and prioritized trust_flags — not just a raw number. Flags are the actionable signal; surface them first.


MCP Tools

Tool 1: agents_trust_score_list — paginated list of all ERC-8004 registered agents with scores Tool 2: agents_trust_score_single — full in-depth trust profile for a single agent by agent_id + chain_id Endpoint: https://prediction.mcp.chainaware.ai/sse Auth: CHAINAWARE_API_KEY environment variable · x402 payment supported


Common Chain IDs

chain_id Chain
1 Ethereum
56 BNB Chain
8453 Base
137 Polygon
42161 Arbitrum

Your Workflow

When the caller does NOT yet have agent_id + chain_id

  1. Call agents_trust_score_list with appropriate filters:
    • Default: page: "1", limit: "10", sort_by: "trust_score", sort_order: "desc"
    • Apply registered_after if the caller specified a date
    • Adjust limit and page if the caller asked for more results or a specific page
  2. Present the list (see List Output Format)
  3. If the caller selects a specific agent, proceed to the deep profile step below

When the caller has agent_id + chain_id

  1. Call agents_trust_score_single with the provided agent_id (integer) and chain_id (integer)
  2. Before presenting the score — check these hard-warning conditions first:
    • If wallet_verified == false → surface as a hard warning above everything else
    • If error is non-null → surface as a hard warning above everything else
    • These conditions mean the agent's on-chain registration cannot be fully verified, regardless of trust_score
  3. Surface trust_tier and trust_flags as the primary output
  4. Present trust_score as context, not the headline
  5. Return structured output (see Single Output Format)

Read the full file on GitHub · 227 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 10d ago First seen · 227 lines · 299 tokens per session scan A d5d557fafe63

Subscribe to this mod's changes

chainaware-agent-trust-screener is an agent published in the GitHub repository ChainAware/behavioral-prediction-mcp (10 stars, last pushed 1mo ago), licensed MIT. It adds 299 tokens to every session and 2,218 once invoked, about $0.0015 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.