Spam Detector Premium Feed Proposal

Name of Premium Feed Spam Detector
Forta Network Bot ID 0xd45f7183783f5893f4b8e187746eaf7294f73a3bb966500d237bd0d5978673fa
About Bot Owner Artem Kovalchuk

I am a web developer by training, and have been an active member of the Forta Community since its inception in the Fall 2021. I participated in several early bot development contests hosted by the Forta Foundation, and won four of them. In the Summer 2022, I was approached by the Foundation about contributing to Forta in a more formal capacity. I entered into an independent contractor agreement with the Foundation and have been participating in core development related activities for the last year, including leading the redesign of the Forta Explorer and Bot Profile pages, and supporting the development of the Attack Detector.

I also developed and maintained one of the most popular bots on the network, the Attack Simulation Bot, which is currently used by the Attack Detector.

Owner Details:
Forta - https://app.forta.network/profile/0x183d13c4fCb5133EA52345A9037C9c25A5Aa139D
LinkedIn - https://www.linkedin.com/in/kovart/
Github - https://github.com/kovart
Twitter - https://twitter.com/kovart_dev

How the Premium Feed works Token Spam has been something that several Forta users, including ZenGo, have highlighted as a problem. Web3 wallets and NFT marketplaces in particular are flooded with spam tokens and NFTs that clutter and degrade the wallet experience. Being able to identify and filter out/hide these assets can both improve the user experience and minimize the risk of scams.

Logic
The Spam Detector is equipped with advanced algorithms that analyze multiple indicators to detect spam tokens. These indicators include token metadata analysis, compliance with declared token standards, distribution rationality, and analysis of creator and recipient behavior. Additionally, the bot has the ability to adjust its assessment during token monitoring, providing accurate and up-to-date detection of spam tokens.

The bot utilizes a modular system of indicators for analyzing tokens. The modules are designed to detect both negative indications of spam and positive indications of a legitimate token. Each module's result is recorded in the overall token analysis context, which can be used by subsequent modules. Additionally, some modules can interrupt the execution of other modules and determine the final evaluation of the token, i.e. `ObservationTimeIsOver` module.

The TokenAnalyzer interprets the final result of the modules and decides whether the token is spam or not. When spam is detected, the bot publishes an alert that includes information about the token and the context of the modules used to evaluate it.

If there are changes during token monitoring, such as when a module finds additional signs of spam, the bot will publish an update alert to indicate the changes. If there is a change that alters the evaluation, such as when the bot no longer identifies the token as spam, it will publish an alert instructing Forta Protocol to remove the labels that identify it as spam.

Supported standards:

  • ERC-20
  • ERC-712
  • ERC-1155

Supported chains (Chain ID):

  • Ethereum (1)
  • BSC (56)
  • Polygon (137)
  • Arbitrum (42161)
  • Optimism (10)
  • Fantom (250)
  • Avalanche (43114)

More details about the Spam Detector can be found in the documentation linked below.

Anticipated target market
  • Web3 wallets
  • NFT marketplaces
  • Large infrastructure providers (Alchemy, Infura)
  • Others
Maintenance/Support I am actively maintaining the bot, testing performance, receiving feedback from the Forta community and making the appropriate upgrades on a weekly basis. This Premium Feed will continue to improve overtime.

I am committing to provide technical support to all subscribers and prospective subscribers via the Feed’s dedicated Discord channel, as well as through direct communication channels.
Fee Recipient 0x183d13c4fCb5133EA52345A9037C9c25A5Aa139D
Price 49 USDC
Performance information The technical performance of the Spam Detector can be monitored on the bot’s page.

The Spam Detector reliably processes incoming alerts as shown on its bot health page:

The bot currently utilizes a sharding configuration of four shards and 3 targets resulting in a reliable deployment of this bot.

Also, the Spam Detector’s current precision is 86.05% (here are more details).

  • Precision is how I quantify the quality of the Spam Detector. It represents the percentage of alerts that are correctly identifying spam tokens (aka a true positive) from 0-100, and calculated on a monthly basis.
  • Precision is calculated using a combination of random sampling and manual review. I take 30 random spam alerts from each token category - ERC20, ERC721 and ERC1155 (90 alerts total). I then review each token’s metadata, transaction activity, etc. to determine if the alert was a true or false positive.
Documentation Documentation
Compatible licensing Forta Bot License 1.0
1 Like

Following the last Forta Council meeting, the Premium Feed proposal submitted by Artem Kovalchuk on the Spam Detector was approved by half of the active Council members in accordance with the Bylaws.

Minutes from Forta Council meeting of July 25th, 2023.

1 Like

After thorough analysis of the user base and the application of the bot, I have decided to adjust the pricing to $599/month.
This change aligns with a strategic shift towards serving the enterprise market more effectively.

2 Likes