Skip to content
Back to search

evmole

EVMole is an open-source library that analyzes deployed EVM bytecode to infer function selectors, argument types, state mutability, persistent and transient storage layout, control flow, and compiler metadata, including for unverified contracts

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked yesterday

Check it deeper on your machine

It has changed since its scan on 18 Nov 2026.

As a Claude Code plugin, install it and then ask Claude Code:

/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/cdump/evmole with scanskillsafety

Or from a terminal, clone it and run it:

git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/cdump/evmole'

It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.

Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/cdump--evmole/badge.svg)](/listings/cdump--evmole)

HTML

<a href="/listings/cdump--evmole"><img src="/api/listings/cdump--evmole/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 535 lines
  1. ![EVMole](./.github/logo.svg)
  2. [![try it online](https://img.shields.io/badge/Try_It_Online-evmole.xyz-brightgreen)](https://evmole.xyz/)
  3. [![npm](https://img.shields.io/npm/v/evmole)](https://www.npmjs.com/package/evmole)
  4. [![Crates.io](https://img.shields.io/crates/v/evmole?color=e9b44f)](https://crates.io/crates/evmole)
  5. [![PyPI](https://img.shields.io/pypi/v/evmole?color=006dad)](https://pypi.org/project/evmole)
  6. [![Go](https://img.shields.io/badge/go-pkg-00ADD8)](https://pkg.go.dev/github.com/cdump/evmole/go)
  7. EVMole is a powerful library that extracts information from Ethereum Virtual Machine (EVM) bytecode, including [function selectors](https://docs.soliditylang.org/en/latest/abi-spec.html#function-selector), arguments, [state mutability](https://docs.soliditylang.org/en/latest/contracts.html#state-mutability), persistent and transient storage layouts, and CBOR metadata, even for unverified contracts.
  8. ## Key Features
  9. - Multi-language support: Available as [JavaScript](#javascript), [Rust](#rust), [Python](#python), and [Go](#go) libraries.
  10. - High accuracy and performance: [Outperforms](#benchmark) existing tools.
  11. - Broad compatibility: Tested with both Solidity and Vyper compiled contracts.
  12. - Lightweight: Clean codebase with minimal external dependencies.
  13. - Unverified contract analysis: Extracts information even from unverified bytecode.
  14. - Selector dispatch classification: Distinguishes normal ABI dispatch from selectors handled by fallback logic.
  15. - CBOR metadata: Extracts string-keyed values from a terminal, length-suffixed CBOR map without assuming a particular compiler.
  16. ## Usage
  17. ### JavaScript
  18. [API documentation](./javascript/#api) and [usage examples](./javascript/#usage) (Node.js, Vite, webpack, Parcel, esbuild)
  19. ```sh
  20. npm i evmole
  21. ```
  22. ```javascript
  23. import { contractInfo } from 'evmole'
  24. const code = '0x6080604052348015600e575f80fd5b50600436106030575f3560e01c80632125b65b146034578063b69ef8a8146044575b5f80fd5b6044603f3660046046565b505050565b005b5f805f606084860312156057575f80fd5b833563ffffffff811681146069575f80fd5b925060208401356001600160a01b03811681146083575f80fd5b915060408401356001600160e01b0381168114609d575f80fd5b80915050925092509256'
  25. console.log( contractInfo(code, {selectors:true, arguments:true, stateMutability:true}) )
  26. // {
  27. // functions: [
  28. // {
  29. // selector: '2125b65b',
  30. // bytecodeOffset: 52,
  31. // dispatch: 'abi',
  32. // arguments: 'uint32,address,uint224',
  33. // stateMutability: 'pure'
  34. // },
  35. // ...
  36. ```
  37. ### Rust
  38. Documentation is available on [docs.rs](https://docs.rs/evmole/latest/evmole/)
  39. ```rust
  40. let code = hex::decode("6080604052348015600e575f80fd5b50600436106030575f3560e01c80632125b65b146034578063b69ef8a8146044575b5f80fd5b6044603f3660046046565b505050565b005b5f805f606084860312156057575f80fd5b833563ffffffff811681146069575f80fd5b925060208401356001600160a01b03811681146083575f80fd5b915060408401356001600160e01b0381168114609d575f80fd5b80915050925092509256").unwrap();
  41. println!("{:?}", evmole::contract_info(
  42. evmole::ContractInfoArgs::new(&code)
  43. .with_selectors()
  44. .with_arguments()
  45. .with_state_mutability()
  46. )
  47. );
  48. // Contract {
  49. // functions: Some([
  50. // Function {
  51. // selector: [33, 37, 182, 91],
  52. // bytecode_offset: 52,
  53. // dispatch: Abi,
  54. // arguments: Some([Uint(32), Address, Uint(224)]),
  55. // state_mutability: Some(Pure)
  56. // },
  57. // ...
  58. ```
  59. ### Python
  60. [API documentation](./python/#api)
  61. ```sh
  62. pip install evmole --upgrade
  63. ```
  64. ```python
  65. from evmole import contract_info
  66. code = '0x6080604052348015600e575f80fd5b50600436106030575f3560e01c80632125b65b146034578063b69ef8a8146044575b5f80fd5b6044603f3660046046565b505050565b005b5f805f606084860312156057575f80fd5b833563ffffffff811681146069575f80fd5b925060208401356001600160a01b03811681146083575f80fd5b915060408401356001600160e01b0381168114609d575f80fd5b80915050925092509256'
  67. print( contract_info(code, selectors=True, arguments=True, state_mutability=True) )
  68. # Contract(
  69. # functions=[
  70. # Function(
  71. # selector=2125b65b,
  72. # bytecode_offset=52,
  73. # dispatch="abi",
  74. # arguments=uint32,address,uint224,
  75. # state_mutability=pure),
  76. # ...
  77. ```
  78. ### Go
  79. [API documentation](./go/#api-reference)
  80. ```sh
  81. go get github.com/cdump/evmole/go
  82. ```
  83. ```go
  84. package main
  85. import (
  86. "context"
  87. "encoding/hex"
  88. "fmt"
  89. "github.com/cdump/evmole/go"
  90. )
  91. func main() {
  92. code, _ := hex.DecodeString("6080604052348015600e575f80fd5b50600436106030575f3560e01c80632125b65b146034578063b69ef8a8146044575b5f80fd5b6044603f3660046046565b505050565b005b5f805f606084860312156057575f80fd5b833563ffffffff811681146069575f80fd5b925060208401356001600160a01b03811681146083575f80fd5b915060408401356001600160e01b0381168114609d575f80fd5b80915050925092509256")
  93. info, _ := evmole.ContractInfo(context.Background(), code, evmole.Options{
  94. Selectors: true,
  95. Arguments: true,
  96. StateMutability: true,
  97. })
  98. for _, fn := range info.Functions {
  99. fmt.Printf("%s: %s @ %d\n", fn.Selector, *fn.Arguments, fn.BytecodeOffset)
  100. }
  101. // 2125b65b: uint32,address,uint224 @ 52
  102. // b69ef8a8: @ 68
  103. }
  104. ```
  105. ### Foundry
  106. <a href="https://getfoundry.sh/">Foundry's cast</a> uses the Rust implementation of EVMole
  107. ```sh
  108. $ cast selectors $(cast code 0xC02aaA39b223FE8D0A0e5C4F27eAD9083C756Cc2)
  109. 0x06fdde03 view
  110. 0x095ea7b3 address,uint256 nonpayable
  111. 0x18160ddd view
  112. 0x23b872dd address,address,uint256 nonpayable
  113. ...
  114. $ cast selectors --resolve $(cast code 0xC02aaA39b223FE8D0A0e5C4F27eAD9083C756Cc2)
  115. 0x06fdde03 view name()
  116. 0x095ea7b3 address,uint256 nonpayable approve(address,uint256)
  117. 0x18160ddd view totalSupply()
  118. 0x23b872dd address,address,uint256 nonpayable transferFrom(address,address,uint256)
  119. ...
  120. ```
  121. ### AI agents
  122. For application code, use one of the language bindings above. For agent-driven
  123. bytecode analysis, choose one integration.
  124. #### JSON CLI
  125. Use for one-off analysis and scripts:
  126. ```bash
  127. npx -y evmole analyze --bytecode 0x...
  128. ```
  129. #### Portable skill
  130. Install routing and interpretation guidance for supported agents:
  131. ```bash
  132. npx skills add cdump/evmole --skill evm-bytecode-analysis -g
  133. ```
  134. #### Local MCP server
  135. Expose EVMole as a typed local tool:
  136. ```bash
  137. npx -y evmole-mcp
  138. ```
  139. All integrations expect deployed/runtime bytecode and run locally without
  140. sending bytecode to an EVMole-operated service. See the
  141. [agent integration guide](./agent/README.md) for setup, schemas,
  142. limitations, and privacy details.
  143. ## Benchmark
  144. ### function selectors
  145. <i>FP/FN</i> - [False Positive/False Negative](https://en.wikipedia.org/wiki/False_positives_and_false_negatives) errors; <b>smaller is better</b>
  146. <table>
  147. <tr>
  148. <td>Dataset</td>
  149. <td></td>
  150. <td><i>evmole</i> <a href="benchmark/providers/evmole-rs/"><i>rs</i></a> · <a href="benchmark/providers/evmole-js/"><i>js</i></a> · <a href="benchmark/providers/evmole-py/"><i>py</i></a> · <a href="benchmark/providers/evmole-go/"><i>go</i></a></td>
  151. <td><a href="benchmark/providers/whatsabi/"><i>whatsabi</i></a></td>
  152. <td><a href="benchmark/providers/sevm/"><i>sevm</i></a></td>
  153. <td><a href="benchmark/providers/evm-hound-rs/"><i>evmhnd</i></a></td>
  154. <td><a href="benchmark/providers/heimdall-rs/"><i>heimdall</i></a></td>
  155. <td><a href="benchmark/providers/gigahorse/"><i>gigahorse</i></a></td>
  156. </tr>
  157. <tr>
  158. <td rowspan="5"><b>coverage2k</b><br><sub>solidity<br><br>2000<br>addresses<br><br>45650<br>functions</sub></td>
  159. <td><i>FP&nbsp;<sub>addr</sub></i></td>
  160. <td>0 🥇</td>
  161. <td>4</td>
  162. <td>1</td>
  163. <td>56</td>
  164. <td>3</td>
  165. <td>2</td>
  166. </tr>
  167. <tr>
  168. <td><i>FN&nbsp;<sub>addr</sub></i></td>
  169. <td>0 🥇</td>
  170. <td>11</td>
  171. <td>0 🥇</td>
  172. <td>224</td>
  173. <td>166</td>
  174. <td>6</td>
  175. </tr>
  176. <tr>
  177. <td><i>FP&nbsp;<sub>func</sub></i></td>
  178. <td>0 🥇</td>
  179. <td>10</td>
  180. <td>10</td>
  181. <td>616</td>
  182. <td>27</td>
  183. <td>11</td>
  184. </tr>
  185. <tr>
  186. <td><i>FN&nbsp;<sub>func</sub></i></td>
  187. <td>0 🥇</td>
  188. <td>145</td>
  189. <td>0 🥇</td>
  190. <td>821</td>
  191. <td>342</td>
  192. <td>211</td>
  193. </tr>
  194. <tr>
  195. <td><i>Time</i></td>
  196. <td>18ms&nbsp;·&nbsp;0.2s&nbsp;·&nbsp;23ms&nbsp;·&nbsp;79ms</td>
  197. <td>2.0s</td>
  198. <td>28s<sup>*</sup></td>
  199. <td>86ms</td>
  200. <td>111s<sup>*</sup></td>
  201. <td>5922s<sup>*</sup></td>
  202. </tr>
  203. <tr><td colspan="8"></td></tr>
  204. <tr>
  205. <td rowspan="5"><b>random10k</b><br><sub>solidity<br><br>10000<br>addresses<br><br>223316<br>functions</sub></td>
  206. <td><i>FP&nbsp;<sub>addr</sub></i></td>
  207. <td>0 🥇</td>
  208. <td>19</td>
  209. <td>16</td>
  210. <td>224</td>
  211. <td>7</td>
  212. <td>21</td>
  213. </tr>
  214. <tr>
  215. <td><i>FN&nbsp;<sub>addr</sub></i></td>
  216. <td>0 🥇</td>
  217. <td>44</td>
  218. <td>1</td>
  219. <td>838</td>
  220. <td>819</td>
  221. <td>16</td>
  222. </tr>
  223. <tr>
  224. <td><i>FP&nbsp;<sub>func</sub></i></td>
  225. <td>0 🥇</td>
  226. <td>65</td>
  227. <td>112</td>
  228. <td>3157</td>
  229. <td>260</td>
  230. <td>92</td>
  231. </tr>
  232. <tr>
  233. <td><i>FN&nbsp;<sub>func</sub></i></td>
  234. <td>0 🥇</td>
  235. <td>173</td>
  236. <td>7</td>
  237. <td>4112</td>
  238. <td>1021</td>
  239. <td>297</td>
  240. </tr>
  241. <tr>
  242. <td><i>Time</i></td>
  243. <td>92ms&nbsp;·&nbsp;0.6s&nbsp;·&nbsp;0.1s&nbsp;·&nbsp;0.5s</td>
  244. <td>7.1s</td>
  245. <td>80s<sup>*</sup></td>
  246. <td>0.4s</td>
  247. <td>533s<sup>*</sup></td>
  248. <td>22870s<sup>*</sup></td>
  249. </tr>
  250. <tr><td colspan="8"></td></tr>
  251. <tr>
  252. <td rowspan="5"><b>coverage1k</b><br><sub>vyper<br><br>1000<br>addresses<br><br>38759<br>functions</sub></td>
  253. <td><i>FP&nbsp;<sub>addr</sub></i></td>
  254. <td>0 🥇</td>
  255. <td>560</td>
  256. <td>0 🥇</td>
  257. <td>2</td>
  258. <td>0 🥇</td>
  259. <td>0 🥇</td>
  260. </tr>
  261. <tr>
  262. <td><i>FN&nbsp;<sub>addr</sub></i></td>
  263. <td>0 🥇</td>
  264. <td>998</td>
  265. <td>788</td>
  266. <td>525</td>
  267. <td>998</td>
  268. <td>927</td>
  269. </tr>
  270. <tr>
  271. <td><i>FP&nbsp;<sub>func</sub></i></td>
  272. <td>0 🥇</td>
  273. <td>560</td>
  274. <td>0 🥇</td>
  275. <td>5</td>
  276. <td>0 🥇</td>
  277. <td>0 🥇</td>
  278. </tr>
  279. <tr>
  280. <td><i>FN&nbsp;<sub>func</sub></i></td>
  281. <td>0 🥇</td>
  282. <td>38759</td>
  283. <td>34077</td>
  284. <td>16218</td>
  285. <td>38759</td>
  286. <td>37858</td>
  287. </tr>
  288. <tr>
  289. <td><i>Time</i></td>
  290. <td>70ms&nbsp;·&nbsp;0.3s&nbsp;·&nbsp;75ms&nbsp;·&nbsp;0.2s</td>
  291. <td>1.9s</td>
  292. <td>5.4s<sup>*</sup></td>
  293. <td>67ms</td>
  294. <td>12s<sup>*</sup></td>
  295. <td>1195s<sup>*</sup></td>
  296. </tr>
  297. </table>
  298. ### function arguments
  299. <i>Errors</i> - when at least 1 inferred argument is incorrect: `(uint256,string)` ≠ `(uint256,bytes)`
  300. <table>
  301. <tr>
  302. <td>Dataset</td>
  303. <td></td>
  304. <td><i>evmole</i> <a href="benchmark/providers/evmole-rs/"><i>rs</i></a> · <a href="benchmark/providers/evmole-js/"><i>js</i></a> · <a href="benchmark/providers/evmole-py/"><i>py</i></a> · <a href="benchmark/providers/evmole-go/"><i>go</i></a></td>
  305. <td><a href="benchmark/providers/heimdall-rs/"><i>heimdall</i></a></td>
  306. </tr>
  307. <tr>
  308. <td rowspan="2"><b>coverage2k</b><br><sub>solidity<br><br>45650<br>functions</sub></td>
  309. <td><i>Errors</i></td>
  310. <td>8.4% 🥇<br><sub>3820</sub></td>
  311. <td>23.3%<br><sub>10643</sub></td>
  312. </tr>
  313. <tr>
  314. <td><i>Time</i></td>
  315. <td>0.6s&nbsp;·&nbsp;1.5s&nbsp;·&nbsp;0.8s&nbsp;·&nbsp;1.8s</td>
  316. <td>111s<sup>*</sup></td>
  317. </tr>
  318. <tr><td colspan="4"></td></tr>
  319. <tr>
  320. <td rowspan="2"><b>random10k</b><br><sub>solidity<br><br>223316<br>functions</sub></td>
  321. <td><i>Errors</i></td>
  322. <td>6.8% 🥇<br><sub>15076</sub></td>
  323. <td>21.4%<br><sub>47878</sub></td>
  324. </tr>
  325. <tr>
  326. <td><i>Time</i></td>
  327. <td>3.0s&nbsp;·&nbsp;6.0s&nbsp;·&nbsp;3.6s&nbsp;·&nbsp;8.5s</td>
  328. <td>511s<sup>*</sup></td>
  329. </tr>
  330. <tr><td colspan="4"></td></tr>
  331. <tr>
  332. <td rowspan="2"><b>coverage1k</b><br><sub>vyper<br><br>38759<br>functions</sub></td>
  333. <td><i>Errors</i></td>
  334. <td>7.4% 🥇<br><sub>2873</sub></td>
  335. <td>100.0%<br><sub>38759</sub></td>
  336. </tr>
  337. <tr>
  338. <td><i>Time</i></td>
  339. <td>0.4s&nbsp;·&nbsp;0.9s&nbsp;·&nbsp;0.5s&nbsp;·&nbsp;1.2s</td>
  340. <td>12s<sup>*</sup></td>
  341. </tr>
  342. </table>
  343. ### function state mutability
  344. <i>Errors</i> - Results are not equal (treating `view` and `pure` as equivalent to `nonpayable`)
  345. <i>Errors strict</i> - Results are strictly unequal (`nonpayable` ≠ `view`). Some ABIs mark `pure`/`view` functions as `nonpayable`, so not all strict errors indicate real issues.
  346. <table>
  347. <tr>
  348. <td>Dataset</td>
  349. <td></td>
  350. <td><i>evmole</i> <a href="benchmark/providers/evmole-rs/"><i>rs</i></a> · <a href="benchmark/providers/evmole-js/"><i>js</i></a> · <a href="benchmark/providers/evmole-py/"><i>py</i></a> · <a href="benchmark/providers/evmole-go/"><i>go</i></a></td>
  351. <td><a href="benchmark/providers/whatsabi/"><i>whatsabi</i></a></td>
  352. <td><a href="benchmark/providers/sevm/"><i>sevm</i></a></td>
  353. <td><a href="benchmark/providers/heimdall-rs/"><i>heimdall</i></a></td>
  354. </tr>
  355. <tr>
  356. <td rowspan="3"><b>coverage2k</b><br><sub>solidity<br><br>45647<br>functions</sub></td>
  357. <td><i>Errors</i></td>
  358. <td>0.0% 🥇<br><sub>18</sub></td>
  359. <td>52.2%<br><sub>23810</sub></td>
  360. <td>11.2%<br><sub>5133</sub></td>
  361. <td>19.6%<br><sub>8951</sub></td>
  362. </tr>
  363. <tr>
  364. <td><i>Errors&nbsp;strict</i></td>
  365. <td>9.0% 🥇<br><sub>4107</sub></td>
  366. <td>70.9%<br><sub>32346</sub></td>
  367. <td>62.3%<br><sub>28429</sub></td>
  368. <td>41.3%<br><sub>18835</sub></td>
  369. </tr>
  370. <tr>
  371. <td><i>Time</i></td>
  372. <td>7.4s&nbsp;·&nbsp;7.6s&nbsp;·&nbsp;7.0s&nbsp;·&nbsp;14s</td>
  373. <td>3.1s</td>
  374. <td>29s<sup>*</sup></td>
  375. <td>112s<sup>*</sup></td>
  376. </tr>
  377. <tr><td colspan="6"></td></tr>
  378. <tr>
  379. <td rowspan="3"><b>random10k</b><br><sub>solidity<br><br>223273<br>functions</sub></td>
  380. <td><i>Errors</i></td>
  381. <td>0.0% 🥇<br><sub>39</sub></td>
  382. <td>48.8%<br><sub>108928</sub></td>
  383. <td>9.3%<br><sub>20713</sub></td>
  384. <td>18.6%<br><sub>41507</sub></td>
  385. </tr>
  386. <tr>
  387. <td><i>Errors&nbsp;strict</i></td>
  388. <td>8.9% 🥇<br><sub>19940</sub></td>
  389. <td>69.9%<br><sub>156081</sub></td>
  390. <td>60.9%<br><sub>136069</sub></td>
  391. <td>40.1%<br><sub>89519</sub></td>
  392. </tr>
  393. <tr>
  394. <td><i>Time</i></td>
  395. <td>35s&nbsp;·&nbsp;34s&nbsp;·&nbsp;33s&nbsp;·&nbsp;67s</td>
  396. <td>12s</td>
  397. <td>81s<sup>*</sup></td>
  398. <td>512s<sup>*</sup></td>
  399. </tr>
  400. <tr><td colspan="6"></td></tr>
  401. <tr>
  402. <td rowspan="3"><b>coverage1k</b><br><sub>vyper<br><br>38278<br>functions</sub></td>
  403. <td><i>Errors</i></td>
  404. <td>0.1% 🥇<br><sub>34</sub></td>
  405. <td>100.0%<br><sub>38278</sub></td>
  406. <td>96.2%<br><sub>36814</sub></td>
  407. <td>100.0%<br><sub>38278</sub></td>
  408. </tr>
  409. <tr>
  410. <td><i>Errors&nbsp;strict</i></td>
  411. <td>1.2% 🥇<br><sub>441</sub></td>
  412. <td>100.0%<br><sub>38278</sub></td>
  413. <td>98.4%<br><sub>37650</sub></td>
  414. <td>100.0%<br><sub>38278</sub></td>
  415. </tr>
  416. <tr>
  417. <td><i>Time</i></td>
  418. <td>34s&nbsp;·&nbsp;31s&nbsp;·&nbsp;36s&nbsp;·&nbsp;71s</td>
  419. <td>2.0s</td>
  420. <td>5.5s<sup>*</sup></td>
  421. <td>12s<sup>*</sup></td>
  422. </tr>
  423. </table>
  424. ### Control Flow Graph
  425. <i>False Negatives</i> - Valid blocks possibly incorrectly marked unreachable by CFG analysis. Lower count usually indicates better precision.
  426. <table>
  427. <tr>
  428. <td></td>
  429. <td><i>evmole</i> <a href="benchmark/providers/evmole-rs/"><i>rs</i></a> · <a href="benchmark/providers/evmole-js/"><i>js</i></a> · <a href="benchmark/providers/evmole-py/"><i>py</i></a> · <a href="benchmark/providers/evmole-go/"><i>go</i></a></td>
  430. <td><a href="benchmark/providers/ethersolve"><i>ethersolve</i></a></td>
  431. <td><a href="benchmark/providers/evm-cfg"><i>evm-cfg</i></a></td>
  432. <td><a href="benchmark/providers/sevm"><i>sevm</i></a></td>
  433. <td><a href="benchmark/providers/heimdall-rs"><i>heimdall-rs</i></a></td>
  434. <td><a href="benchmark/providers/evm-cfg-builder"><i>evm-cfg-builder</i></a></td>
  435. </tr>
  436. <tr>
  437. <td><i>Basic&nbsp;Blocks</i></td>
  438. <td>92.8% 🥇<br><sub>483212</sub></td>
  439. <td>52.5%<br><sub>273518</sub></td>
  440. <td>58.6%<br><sub>305248</sub></td>
  441. <td>37.3%<br><sub>194368</sub></td>
  442. <td>32.6%<br><sub>169980</sub></td>
  443. <td>14.5%<br><sub>75383</sub></td>
  444. </tr>
  445. <tr>
  446. <td><i>False&nbsp;Negatives</i></td>
  447. <td>7.2% 🥇<br><sub>37496</sub></td>
  448. <td>47.5%<br><sub>247190</sub></td>
  449. <td>41.4%<br><sub>215460</sub></td>
  450. <td>62.7%<br><sub>326340</sub></td>
  451. <td>67.4%<br><sub>350728</sub></td>
  452. <td>85.5%<br><sub>445325</sub></td>
  453. </tr>
  454. <tr>
  455. <td><i>Time</i></td>
  456. <td>13s&nbsp;·&nbsp;24s&nbsp;·&nbsp;11s&nbsp;·&nbsp;47s</td>
  457. <td>888s</td>
  458. <td>36s</td>
  459. <td>9.8s</td>
  460. <td>20s</td>
  461. <td>359s</td>
  462. </tr>
  463. </table>
  464. dataset flow-challenge500, 500 contracts, 520,708 blocks
  465. ### notes
  466. See [benchmark/README.md](./benchmark/) for the methodology and commands to reproduce these results
  467. <i>versions: evmole v0.9.5; <a href="https://github.com/shazow/whatsabi">whatsabi</a> v0.25.0; <a href="https://github.com/acuarica/evm">sevm</a> v0.7.4; <a href="https://github.com/g00dv1n/evm-hound-rs">evm-hound-rs</a> v0.1.4; <a href="https://github.com/Jon-Becker/heimdall-rs">heimdall-rs</a> v0.9.3; <a href="https://github.com/nevillegrech/gigahorse-toolchain/tree/1b06cef82f0a29790e87f576eba6d6d793170d70">gigahorse</a> commit 1b06cef</i>
  468. <sup>*</sup>: <b>sevm</b>, <b>heimdall-rs</b>, and <b>gigahorse</b> are full decompilers, not limited to extracting function selectors
  469. ## How it works
  470. EVMole uses symbolic execution with a custom EVM implementation to trace how CALLDATA flows through the bytecode:
  471. This approach is more accurate than static pattern matching because it follows the actual execution paths the EVM would take, correctly handling complex dispatchers, proxy patterns, and compiler-specific optimizations from both Solidity and Vyper.
  472. ## Talks
  473. - [EVMole: function selectors and arguments from bytecode](https://www.youtube.com/watch?v=l0udabGej54) - BlockSplit 2024
  474. - [EVMole: function selectors and arguments from bytecode](https://ethcc.io/archives/evmole-function-selectors-and-arguments-from-bytecode) - EthCC 2024
  475. - [Reconstructing Control Flow Graphs from EVM Bytecode](https://www.youtube.com/watch?v=1Xd6PhEHMHM) - ETHTaipei 2025
  476. - [Reconstructing Control Flow Graphs from EVM Bytecode: Faster, Better, Stronger](https://www.youtube.com/watch?v=UL6-3EZbv3E) - EthCC 2025
  477. ## License
  478. MIT

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.