Skip to content

A smart contract security search utility

Notifications You must be signed in to change notification settings

baoanh1310/Masamune

 
 

Repository files navigation

The smart contract security search utility tool.

What is Masamune?

Masamune is a search utility tool that allows you to search for smart contract security vulnerabilities, from a curated list of sources.

To access Masamune, visit masamune.app.

How does it work?

Masamune V1.

The search utility is powered by Lunr.js, a full-text search library for the browser.

We have developed custom scrapers for each data source, which are run periodically to retrieve the latest data. You can find the scrapers in the scrapers directory.

The data is stored within the results directory; for each of the queries, a pattern match is tried against the data, and the results are displayed.

To build locally, just open index.html using a live server, eg. this extension for VSCode.

Masamune V2.

The second iteration of Masamune is currently under development, however a beta version will be available at masamune.app soon. The new version will be powered by OpenAI's Embeddings API, which will allow for more advanced search queries, as well as more context aware search results.

Retrieving the data

Currently, Masamune works on the following data sources:

  1. Code4rena findings.
  2. Immunefi bugfixes.
  3. DeFi Hacks Analysis.
  4. Zellic audits.
  5. yAudit findings.
  6. Trail of Bits audits
  7. Various Gitbooks, such as the Layer Zero Docs, Curve Finance Docs, MEV Wiki, etc.

Creating a custom scraper

To add a new data source, a custom scraper for the report format must be created. Existing examples are found in the scrapers directory.

Scrapers have 2 main functions. The extract_finding() function parses the report files and stores the stored output in a text file in the findings_newupdate directory. The jsonify_findings() function parses this text file and outputs a JSON file stored in the results directory. The format of the JSON data is found in other scraper files but generally follows this format:

        "title": "MobileCoin Foundation could infer token IDs in certain scenarios",
        "labels": [
            "Trail of Bits",
            "Severity: Informational",
            "Difficulty: High",
            "Type: Data Exposure",
        ]
        "body": ...

For Masamune V1, after the new scraper is created and the JSON output is stored in the results directory, the JSON file must be added to the dataset variable in logic.js to be included in the search results.

About

A smart contract security search utility

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 69.1%
  • JavaScript 16.1%
  • HTML 7.9%
  • CSS 3.9%
  • SCSS 2.3%
  • Dockerfile 0.7%