zkvm-evaluator
Trustless ERC-8183 job evaluation — run Client's verification program inside a zkVM with ZK proof.
Best use case
zkvm-evaluator is best used when you need a repeatable AI agent workflow instead of a one-off prompt. It is especially useful for teams working in multi. Trustless ERC-8183 job evaluation — run Client's verification program inside a zkVM with ZK proof.
Trustless ERC-8183 job evaluation — run Client's verification program inside a zkVM with ZK proof.
Users should expect a more consistent workflow output, faster repeated execution, and less time spent rewriting prompts from scratch.
Practical example
Example input
Use the "zkvm-evaluator" skill to help with this workflow task. Context: Trustless ERC-8183 job evaluation — run Client's verification program inside a zkVM with ZK proof.
Example output
A structured workflow result with clearer steps, more consistent formatting, and an output that is easier to reuse in the next run.
When to use this skill
- Use this skill when you want a reusable workflow rather than writing the same prompt again and again.
When not to use this skill
- Do not use this when you only need a one-off answer and do not need a reusable workflow.
- Do not use it if you cannot install or maintain the related files, repository context, or supporting tools.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/zkvm-evaluator/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How zkvm-evaluator Compares
| Feature / Agent | zkvm-evaluator | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Trustless ERC-8183 job evaluation — run Client's verification program inside a zkVM with ZK proof.
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
# zkVM Evaluator
Trustless evaluation for ERC-8183 Agentic Commerce jobs. The Client provides a verification program, the Evaluator runs it on the Provider's deliverable inside a zkVM, and the zkVM generates a mathematical proof that the program actually executed and produced the claimed result.
**No trust in the evaluator needed.** The proof is verifiable on-chain by anyone.
## Tools
- `zkvm_evaluate_job` — Run verification program on deliverable, generate ZK proof, settle job on-chain
- `zkvm_status` — Check which zkVM backends are installed (SP1, RISC Zero, native)
## How It Works
```
Client creates Job:
- description: "Write an ERC-20 contract"
- verification program: test_suite.elf → IPFS (QmXyz...)
- evaluator: zkVM evaluator contract
Provider submits:
- deliverable: MyToken.sol → IPFS (QmAbc...)
Evaluator runs:
zkvm_evaluate_job(
jobId: "42",
programRef: "QmXyz...", ← Client's test suite
deliverableRef: "QmAbc..." ← Provider's code
)
Inside zkVM:
1. Load test_suite.elf (RISC-V binary)
2. Feed MyToken.sol as stdin
3. Execute: test_suite(MyToken.sol)
4. Result: exit code 0 (all tests pass)
5. Generate ZK proof of execution
On-chain:
proof proves "program QmXyz on input QmAbc exited with code 0"
→ complete(jobId=42, reasonHash=proof_hash)
→ funds released to Provider
```
## Verification Program Requirements
The verification program is a compiled RISC-V ELF binary that:
- Reads the deliverable from **stdin**
- Reads optional context from env vars (`ZKVM_JOB_DESCRIPTION`, `ZKVM_DELIVERABLE_HASH`)
- Outputs evaluation details to **stdout**
- Exits with code **0** for pass, **non-zero** for fail
Example (Rust, compiled for RISC-V):
```rust
use std::io::Read;
fn main() {
let mut deliverable = String::new();
std::io::stdin().read_to_string(&mut deliverable).unwrap();
// Check deliverable meets requirements
if deliverable.contains("function transfer")
&& deliverable.contains("function approve")
&& deliverable.contains("event Transfer")
{
println!("PASS: ERC-20 interface complete");
std::process::exit(0);
} else {
println!("FAIL: Missing required ERC-20 functions");
std::process::exit(1);
}
}
```
## Backends
| Backend | ZK Proof | Speed | Install |
| ------------------ | ------------------- | ------------------------------------ | -------------------------------------------------------------- |
| **SP1** (Succinct) | Yes (Groth16/PLONK) | Fastest | `curl -L https://sp1.succinct.xyz \| bash && sp1up` |
| **RISC Zero** | Yes (Groth16) | Fast, Bonsai remote prover available | `curl -L https://risczero.com/install \| bash && rzup install` |
| **Native** | No (dev only) | Instant | Always available |
Auto-detection: the tool picks the best available backend. Use `zkvm_status` to check.
## What the Proof Guarantees
```
Public inputs (anyone can see):
- Program hash: SHA-256 of the verification program
- Deliverable hash: SHA-256 of the deliverable
- Exit code: 0 (pass) or non-zero (fail)
The proof mathematically guarantees:
✅ This specific program ran on this specific deliverable
✅ The exit code is what the program actually produced
✅ No one tampered with the execution
The proof does NOT reveal:
❌ The deliverable content (only its hash)
❌ Internal program state during execution
```
## On-Chain Cost
| Proof System | Verification Gas | Cost on Base L2 |
| ------------ | ---------------- | --------------- |
| Groth16 | ~200k gas | ~$0.004 |
| PLONK | ~300k gas | ~$0.006 |
## Integration with TEE
For maximum security, run the zkVM evaluator inside a TEE enclave:
- TEE: proves the evaluator code wasn't tampered with
- ZK: proves the verification program produced the claimed result
- Combined: hardware + mathematical proof — strongest guaranteeRelated Skills
github
GitHub operations via `gh` CLI: issues, PRs, CI runs, code review, API queries. Use when: (1) checking PR status or CI, (2) creating/commenting on issues, (3) listing/filtering PRs or issues, (4) viewing run logs. NOT for: complex web UI interactions requiring manual browser flows (use browser tooling when available), bulk operations across many repos (script with gh api), or when gh auth is not configured.
gifgrep
Search GIF providers with CLI/TUI, download results, and extract stills/sheets.
xurl
A CLI tool for making authenticated requests to the X (Twitter) API. Use this skill when you need to post tweets, reply, quote, search, read posts, manage followers, send DMs, upload media, or interact with any X API v2 endpoint.
whale-watcher
Monitor large transactions and whale movements on-chain.
weather
Get current weather and forecasts via wttr.in or Open-Meteo. Use when: user asks about weather, temperature, or forecasts for any location. NOT for: historical weather data, severe weather alerts, or detailed meteorological analysis. No API key needed.
wallet-manager
Create, import, and manage blockchain wallets securely.
wacli
Send WhatsApp messages to other people or search/sync WhatsApp history via the wacli CLI (not for normal user chats).
voice-call
Start voice calls via the OpenClaw voice-call plugin.
video-frames
Extract frames or short clips from videos using ffmpeg.
trello
Manage Trello boards, lists, and cards via the Trello REST API.
token-swap
Execute token swaps on Uniswap/PancakeSwap across multiple EVM chains.
tmux
Remote-control tmux sessions for interactive CLIs by sending keystrokes and scraping pane output.