Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →ERC-8004 Trustless Agents — on-chain agent identity + reputation. Resolve an agent by id (Identity Registry ERC-721 → owner + AgentCard), list the canonical registry addresses, and generate a spec-compliant agent registration card (registration-v1, with x402 support + trust models). Custody-free, read + scaffolding. Triggers: ERC-8004, trustless agent, agent identity, agent registry, AgentCard, agent reputation, on-chain agent, A2A, agent discovery, agent card, 8004.
.claude/skills/internet-court-trustless-agents/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | -56% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -36% | 0% |
| case-16 | ✗→✓ | ▲ Improved | -61% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -50% | 0% |
| case-21 | ✗→✓ | ▲ Improved | -48% | 0% |
ERC-8004 (draft EIP) gives AI agents a portable on-chain identity + reputation: the Identity Registry is an ERC-721 ("AgentIdentity") deployed as a singleton at the vanity address 0x8004A169… across many chains (verified on Base mainnet). An agent's tokenURI points to an AgentCard describing what it does, its endpoints (A2A/MCP), trust models, and x402 support.
| Tool | Purpose | |---|---| | chaingpt_erc8004_resolve_agent | Resolve agentId → owner + decoded AgentCard (handles data: and https/ipfs tokenURIs). Read-only. | | chaingpt_erc8004_registries | The canonical Identity + Reputation registry addresses and the chains they're on. | | chaingpt_erc8004_agentcard | Generate a registration-v1 AgentCard JSON (name, services, supportedTrust, x402Support) to host at the tokenURI / /.well-known/agent-card.json. Optionally emits the data: URI form. |
chaingpt_erc8004_resolve_agent agentId=0 chain=base → the Genesis Agent.
Read + scaffolding are shipped and verified. The write path (register / giveFeedback / validation) is intentionally deferred: ERC-8004 is a draft whose write ABIs (especially the Validation Registry) are still being revised, so we don't ship unverified identity/reputation-mutating calldata. Generate the AgentCard here, then register via the project's reference contracts once the write ABI is final.
Reads use a public-RPC fallback chain; set BASE_RPC_URL for reliability. 0 ChainGPT credits.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-10 | pass→pass | 9,140 | 1,743 | -81% | 1 | 1 | 0% | 1,661 | 705 | -58% | 0 | 0 | — |
case-01 | fail→fail | 8,307 | 4,479 | -46% | 1 | 1 | 0% | 1,705 | 681 | -60% | 0 | 0 | — |
case-02 | fail→fail | 24,928 | 7,293 | -71% | 1 | 1 | 0% | 1,161 | 1,061 | -9% | 0 | 0 | — |
case-03 | fail→fail | 9,664 | 2,346 | -76% | 1 | 1 | 0% | 1,621 | 582 | -64% | 0 | 0 | — |
case-04 | pass→pass | 7,375 | 4,182 | -43% | 1 | 1 | 0% | 1,350 | 1,063 | -21% | 0 | 0 | — |
case-05 | fail→fail | 5,754 | 4,660 | -19% | 1 | 1 | 0% | 1,082 | 853 | -21% | 0 | 0 | — |
case-15 | fail→fail | 10,310 | 4,197 | -59% | 1 | 1 | 0% | 1,868 | 632 | -66% | 0 | 0 | — |
case-06 | fail→fail | 7,147 | 4,898 | -31% | 1 | 1 | 0% | 1,242 | 472 | -62% | 0 | 0 | — |
case-07 | fail→fail | 13,789 | 1,926 | -86% | 1 | 1 | 0% | 2,586 | 612 | -76% | 0 | 0 | — |
case-08 | fail→pass | 10,745 | 5,375 | -50% | 1 | 1 | 0% | 2,311 | 1,028 | -56% | 0 | 0 | — |
case-09 | fail→fail | 17,958 | 3,385 | -81% | 1 | 1 | 0% | 4,337 | 924 | -79% | 0 | 0 | — |
case-11 | pass→pass | 8,038 | 1,583 | -80% | 1 | 1 | 0% | 1,392 | 678 | -51% | 0 | 0 | — |
case-12 | pass→pass | 6,573 | 1,881 | -71% | 1 | 1 | 0% | 1,204 | 771 | -36% | 0 | 0 | — |
case-13 | pass→pass | 6,335 | 2,620 | -59% | 1 | 1 | 0% | 1,214 | 925 | -24% | 0 | 0 | — |
case-14 | fail→pass | 10,247 | 3,915 | -62% | 1 | 1 | 0% | 1,795 | 1,147 | -36% | 0 | 0 | — |
case-16 | fail→pass | 11,093 | 1,622 | -85% | 1 | 1 | 0% | 1,948 | 751 | -61% | 0 | 0 | — |
case-17 | fail→pass | 8,185 | 4,233 | -48% | 1 | 1 | 0% | 1,544 | 767 | -50% | 0 | 0 | — |
case-18 | fail→fail | 11,741 | 4,936 | -58% | 1 | 1 | 0% | 2,129 | 791 | -63% | 0 | 0 | — |
case-19 | fail→fail | 10,932 | 2,482 | -77% | 1 | 1 | 0% | 2,224 | 738 | -67% | 0 | 0 | — |
case-20 | pass→pass | 11,082 | 3,728 | -66% | 1 | 1 | 0% | 1,825 | 1,076 | -41% | 0 | 0 | — |
case-21 | fail→pass | 13,573 | 5,256 | -61% | 1 | 1 | 0% | 2,574 | 1,337 | -48% | 0 | 0 | — |
case-22 | pass→pass | 7,838 | 1,768 | -77% | 1 | 1 | 0% | 1,484 | 748 | -50% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 15 counted toward the lift figure. The other 7 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +23 percentage points is the difference between those two pass rates over the 15 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.