🏛️ ERC-8004 Reputation Provider — Live on Base

We See What Agents
Think Before They Act

The reference Reputation Registry for ERC-8004 Trustless Agents — analyzing CoT traces, tool calls, and goal drift to predict blow-ups before wallet tx. Not credit scores — forensic autopsies. EAS-attested on Base.

View Pricing View Live Data
Claims Tracked
Value Lost
Agents Tracked
10
Behavioral Archetypes
ERC-8004
Reputation Provider
Compatible with ERC-8004 · Payments via x402 · Attested on Base
🧠 THE MOAT

Agentic Internals Forensics

While others check wallet history (reactive), we analyze reasoning patterns (predictive)

🧠

CoT Trace Analysis

Detect exploit patterns in chain-of-thought reasoning before execution

🔄

Goal Drift Detection

Flag when objectives shift from "safe swap" to "max extract"

🔧

Tool-Call Graphs

Identify looping denial, gas griefing, infinite retry patterns

🔓

Jailbreak Resistance

Test if agent succumbs to prompt injection attacks

📝

Self-Modification

Detect unauthorized prompt/config changes by the agent itself

⚠️

Reckless Planning

Catch unlimited approvals, no stop-loss, high leverage plans

SAMPLE FORENSIC AUTOPSY Agent: ****_trader_42
D
Risk Score: 42/100
Recurrence: 73% in 30 days

Key Behavioral Archetypes:

72% Goal_Drift_Hijack "Objective shifted from 'safe swap' to 'extract maximum' in step 17"
58% Reckless_Planning "Planning unlimited token approvals without slippage protection"
41% Tool_Looping_Denial "Tool 'execute_swap' failed 7 times with retry escalation"
AlliGo Forensics: D – Goal Drift Detected
Includes: CoT traces, tool graphs, memory analysis, recurrence forecast
// Analyze agentic internals
POST https://alligo-production.up.railway.app/api/forensics/agentic
// Body: { agent_handle, direct_agentic_data: { cot_trace, tool_calls, ... } }

// Response
{
  "overall_risk_score": 42,
  "grade": "D",
  "behavioral_archetypes": [
    { "name": "Goal_Drift_Hijack", "probability": 72, "severity": "critical" },
    { "name": "Reckless_Planning", "probability": 58, "severity": "high" }
  ],
  "recurrence_forecast": { "risk_level": "critical", "probability": 73 },
  "badge_suggestion": "AlliGo Forensics: D – Goal Drift Detected"
}
⚠ LIVE RISK ALERTS

Pre-Mortem Predictions

Agents flagged before the incident. Every alert is timestamped onchain via EAS — verifiable proof of prediction.

⚠️ Predictive signals based on agentic internals forensics. Not financial advice. Past predictions do not guarantee future outcomes. AlliGo is not liable for losses. All alerts ≥80% confidence only.
Active Flags
Confirmed
Total Predictions
Loading risk alerts…

🔒 Full internals — archetype confidence breakdown, EAS proof links, and raw forensics signals — require API access.

Pro API key — $99 / mo

Ready to integrate?

Choose your plan above and request access — API key issued within 24h.

Payment via USDC on Base.

🎂 ERC-8004 GENESIS MONTH — MARCH 2026

AlliGo is the Reputation Layer for ERC-8004

ERC-8004 (Trustless Agents) defines three on-chain registries: Identity, Reputation, and Validation. AlliGo is the reference implementation of the Reputation Registry — behavioral forensics, EAS-attested on Base, queryable via x402.

🆔
Identity Registry
ERC-721 per agent. We index and score any registered agent automatically.
🧠
Reputation Registry ← AlliGo
10 behavioral archetypes. 96 incidents. 92% prediction accuracy. EAS-attested on Base.
Validation Registry
Our forensics signals feed into independent validators and zkML verifiers.
Read EIP-8004 → Get Reputation API Access →

💸 Agent Failure Leaderboard

Real incidents. Real money lost. Live from the AlliGo database — updated by autonomous swarm agents.

Loading leaderboard…

🤖 Daydreams Commerce Harness

LIVE · ERC-8004

AlliGo is scoring every agent on the Daydreams TaskMarket in real-time — the missing Reputation primitive in the Commerce Harness. Powered by the same forensic engine that attests on Base.

Agents Scored
High Risk
Avg Risk Score
Base
Network
Loading Daydreams agents…

Commerce Harness: Wallet · Tasks · Messaging · Search · Identity · Reputation ← AlliGo · Email

The Problem: Flying Blind

When AI agents fail, lose money, or cause damage — there's no record. You're trusting them with zero data.

💸

$250K Lobstar Wilde

OpenAI dev's agent sent 52M tokens to a stranger. State management failure. No record until we logged it.

🔓

$25M Arup Deepfake

AI deepfakes impersonated CFO, authorized $25M transfer. No verification system caught it.

📉

$20M Portfolio Collapse

AI-managed portfolio lost $20.4M on agent tokens. No stop-losses, no position limits.

$7M KiloEx Exploit

Flash loan exploit drained $7M. Oracle manipulation. Common pattern, no centralized tracking.

The pattern: Every failure could have been prevented with better information.

The Solution: AlliGo Risk Scores

The first Agent Claims Registry — real failure data, real risk scores, real protection.

Agent Fails
Claim Logged
Score Calculated
You Get Alerted
📊

Risk Scores (0-100)

Every agent gets a risk score and grade (A-F). Check before you transact. 3-second API call.

🔍

Claims Database

Verified incident records: amount lost, root cause, resolution. Real data from real failures.

Real-time Alerts

Telegram alerts when new claims are filed. Get notified before it's too late.

🏷️

Trust Badges

"AlliGo Certified" badges for clean agents. Embed in your site, verify via API.

// Check agent risk in 3 seconds
GET https://alligo-production.up.railway.app/api/public/agents/eliza_trader_001/score

// Response
{
  "riskScore": 72.5,
  "grade": "B",
  "totalClaims": 2,
  "totalValueLost": 57000,
  "summary": "Good track record. 2 claims, $57K total loss."
}

// Block high-risk agents
if (score.grade === 'F' || score.grade === 'D') {
  throw new Error("Agent not trusted. Check AlliGo for details.");
}

How It Works

Integrate in under 5 minutes

1

Get API Key

Sign up free. Get instant access to 100 requests/day. No credit card needed.

2

Check Agent Score

One API call. Get risk score, grade, claims history, and summary. 3-second response.

3

Protect Your Users

Block high-risk agents. Display trust badges. Get alerts on new claims.

Choose Your Plan

Start free, scale as you grow. All plans include full API access.

STARTER
Free
$0/month
Perfect for trying out AlliGo
  • 100 API requests/day
  • Full claims database
  • Agent risk scores
  • Badge embedding
  • Telegram alerts
  • Webhooks
FOR PLATFORMS
Enterprise
Custom
For DeFi protocols & platforms
  • Unlimited API requests
  • Private claims database
  • Custom risk models
  • White-label badges
  • Dedicated support
  • Insurance partnerships

Tracking failures from agents built on:

Eliza
Virtuals
ai16z
Hyperliquid
Base
Solana

Start Protecting Your Users Today

Join platforms using AlliGo to verify agent trustworthiness. Free tier available — no credit card required.

Join Telegram Alerts

Free forever tier • ✅ Setup in 5 minutes • ✅ Cancel anytime

ERC-8004 Reputation Provider · x402 native · EAS attested on Base