Specialised Agents¶

Q: How does the GameFi screener distinguish human players from bots?

It uses predictive_behaviour signals: bots typically show machine-regular transaction timing, uniform interaction patterns, and zero-diversity engagement. Human players exhibit organic timing variation, protocol diversity, and non-repetitive activity. The agent classifies players as Genuine / Suspected Bot / Bot Farm / Fraudulent based on these signals.

Key Takeaways
- Two vertical-specific agents: chainaware-gamefi-screener for P2E bot detection and player authenticity, chainaware-governance-screener for DAO Sybil attack detection and vote integrity
- chainaware-gamefi-screener classifies players as Genuine / Suspected Bot / Bot Farm / Fraudulent - protecting in-game economies and tournament prize pools
- chainaware-governance-screener detects Sybil clusters, assigns governance quality scores, and recommends voting weight multipliers (Full / Reduced / Excluded)
- Both agents combine predictive_behaviour and predictive_fraud for authentic vs. scripted activity distinction

Two agents designed for specific Web3 verticals: GameFi platforms and DAO governance systems. Both apply ChainAware's behavioural and fraud prediction to the unique requirements of their respective use cases.

Setup required: For AI Agents - MCP registration and agent installation.

chainaware-gamefi-screener¶

Role: Screens players and wallets interacting with GameFi platforms for bot activity, multi-accounting, and fraud.

What it does: GameFi platforms face specific threats: bot farms harvesting in-game rewards, Sybil accounts claiming multiple airdrops or tournament prizes, and fraudulent asset transfers. This agent evaluates player wallets using predictive_behaviour (bot-like vs. human engagement patterns) and predictive_fraud (fraud history) to classify players and protect game economies.

Tools used: predictive_behaviour, predictive_fraud
Model: Claude Haiku 4.5

Example invocation:

@chainaware-gamefi-screener Screen this player wallet before they claim tournament rewards: 0xPLAYER...WALLET

Output includes:
- Player authenticity classification (Genuine Player / Suspected Bot / Bot Farm / Fraudulent)
- Behavioural signals driving the classification (transaction timing, pattern regularity, multi-wallet indicators)
- Fraud risk score
- Reward eligibility recommendation (Allow / Review / Deny)
- Specific flags for game economy protection (multi-accounting suspicion, bot farming patterns)

Common use cases:
- Tournament reward eligibility verification
- NFT airdrop Sybil prevention for GameFi communities
- Player KYC for regulated gaming jurisdictions
- In-game asset marketplace fraud prevention
- Anti-bot protection for play-to-earn mechanics

chainaware-governance-screener¶

Role: Screens DAO governance participants for Sybil attacks, vote manipulation, and low-quality participation.

What it does: DAO governance is vulnerable to Sybil attacks (one entity controlling many wallets to amplify vote weight) and to voter apathy or mercenary voting (selling governance influence). This agent assesses participating wallets using predictive_behaviour (genuine DAO engagement, voting history patterns, wallet tenure) and predictive_fraud (Sybil and fraud signals) to produce a governance quality score.

Tools used: predictive_behaviour, predictive_fraud
Model: Claude Haiku 4.5

Example invocation:

@chainaware-governance-screener Screen these governance voters for Sybil activity: [0xVOTER1, 0xVOTER2, 0xVOTER3]

Output includes:
- Per-wallet governance quality score
- Sybil risk classification (Genuine Participant / Suspicious / Sybil Cluster Indicator)
- Fraud risk overlay
- Voting weight recommendation (Full / Reduced / Excluded)
- Cluster analysis (wallets with correlated behaviour flagged as potential Sybil groups)
- Overall governance health summary for the proposal or snapshot

Common use cases:
- Snapshot vote integrity verification before results are actioned
- Governance token airdrop Sybil prevention
- Delegated voting eligibility screening
- Treasury proposal voting quality assessment
- On-chain governance audit for regulatory or community transparency

Frequently Asked Questions¶

How does the GameFi screener distinguish human players from bots?
It uses predictive_behaviour signals: bots typically show machine-regular transaction timing, uniform interaction patterns, and zero-diversity engagement. Human players exhibit organic timing variation, protocol diversity, and non-repetitive activity. The agent classifies players as Genuine / Suspected Bot / Bot Farm / Fraudulent based on these signals.

Can the governance screener catch Sybil attacks before a vote is completed?
Yes - it is designed to be called before results are actioned (for example, before a Snapshot vote is finalised). It identifies correlated wallet behaviour across the voter list, flags potential Sybil clusters, and recommends voting weight adjustments (Full / Reduced / Excluded) so governance outcomes can be verified before implementation.

Does a "Reduced" or "Excluded" recommendation change the smart contract?
No. The agent returns a recommendation only - it does not modify on-chain contracts. The output is a structured report with per-wallet classifications and suggested weight multipliers that your protocol's governance process or admin team can act on.

Specialised Agents¶

chainaware-gamefi-screener¶

chainaware-governance-screener¶

Frequently Asked Questions¶

Further Reading¶