Case notes
How the score works
Every score on this platform comes from one deterministic function over seven weighted categories. The same inputs always produce the same output, and every number is shown alongside the signals that produced it.
100 points total
Category weights
What each category measures and how much it can contribute to the rug risk score.
Contract Control
20%Powers the deployer still holds over the token right now: minting new supply, freezing holder accounts, and withdrawing pooled liquidity. Read from present state, so it works on a token minutes old.
Developer History
18%What happened to the tokens this wallet deployed before. The strongest single predictor we track.
Supply Concentration
16%How much of the total supply sits in wallets clustered as insiders rather than open-market buyers.
Creator Selling
15%How much of their own allocation the deployer has sold, and how quickly they sold it.
Wallet Connections
13%Whether the deployer's funding graph touches wallets with their own collapse history.
Liquidity Behavior
12%Whether pooled liquidity has been withdrawn, and how much of it is left for holders to exit into.
Launch Bundling
4%Supply acquired in the launch block by wallets that appear coordinated rather than independent.
Social Reputation
2%Age and activity of associated accounts. Freshly created accounts carry no reputational cost to abandon.
0 to 100
Risk bands
Where the thresholds sit. These are the same values the score rulers mark across the platform.
- 0–29Low Risk
- 30–54Moderate Risk
- 55–77High Risk
- 78–100Critical Risk
Run through the live engine
Worked example
A hypothetical launch, scored by the same function that scores every token on the platform. These numbers are computed on page load, not written by hand.
Inputs
- Previous launches
- 9
- Previous collapses
- 6
- Creator sold
- 78%
- Insider supply
- 31%
- Bundled supply
- 24%
- Connected risky wallets
- 3
- Liquidity removed
- 55%
- Social account age
- 4 days
Category contributions
- Contract Control0.0 / 20
- Developer History12.9 / 18
- Supply Concentration11.0 / 16
- Creator Selling13.5 / 15
- Wallet Connections8.6 / 13
- Liquidity Behavior6.6 / 12
- Launch Bundling2.4 / 4
- Social Reputation1.5 / 2
Generated reasoning
- Creator sold 78% of holdings within the first day.
- Developer launched nine previous tokens, and six of them collapsed by more than 90% within 24 hours.
- Wallets clustered with the deployer hold 31% of total supply.
- Three wallets in this deployer's funding graph have a collapse history of their own. Initial buyers were funded from the same source.
- 55% of pooled liquidity has been withdrawn since launch.
- 24% of supply was bought in the launch block by wallets funded from the same source.
- The associated account was created four days before launch. This association is unverified.
Trust is not the inverse of risk
Developer trust score
Rug risk asks what this launch looks like. Trust asks what happened to everything this wallet deployed before — they are different questions and are calculated separately.
Trust starts from the share of a deployer’s resolved launches that did not collapse, then subtracts for selling behaviour, liquidity withdrawal and links to wallets with their own collapse history. Launches that are still trading do not count as survivals yet — otherwise a wallet whose first token simply has not failed would score as highly as one with years of clean history.
Scores are then pulled toward the 50-point midpoint in proportion to how little evidence exists. One resolved launch is weak evidence; twelve is strong. This is why a brand-new wallet sits near 50 rather than at either extreme, and why the confidence label matters as much as the number.
Read this part
Limitations
Where this analysis is weak, stated plainly.
Scores are risk indicators. They are not proof of fraud, and they are not accusations against any individual. Blockchain clustering produces false positives. Always conduct your own research.
Scores are indicators, not proof
A high rug risk score means a launch shares structural features with launches that have collapsed before. It is not a finding of fraud, a legal conclusion, or a statement about anyone's intent. Tokens with high scores sometimes do fine, and tokens with low scores sometimes collapse.
Wallet clustering produces false positives
Links between wallets are inferred from funding flows. Centralised exchanges, bridges and aggregators sit in the funding path of enormous numbers of unrelated wallets, and a naive walk through that graph will connect people who have never interacted. Every link carries its own confidence value — treat anything below 50 as weak.
Creator attribution is imperfect
The wallet that signs a deployment is not always the party that benefits from it. Attribution follows funding one or two hops back where the trail is clear, and stops where it is not. A deployer who uses a fresh wallet each time will have a shorter history than their actual record.
Social associations are unverified
Handles listed in token metadata are self-declared. Anyone can put any account in that field, including one belonging to someone with no connection to the project. Community-submitted links are also unverified. Nothing on this platform confirms that an account belongs to a deployer.
A clean record is not a guarantee
Every wallet on our repeat-offender board had no history at all on its first day. Absence of a bad record means we have not observed one yet, which is a weaker statement than it looks. Confidence is reported alongside every score for exactly this reason.
Do your own research
This platform is one input. It cannot see intent, off-chain agreements, or anything that has not happened yet. Verify independently before making any decision that matters to you.
Disclaimer
OnlyRugs provides automated blockchain risk analysis and community-submitted information. Scores and wallet associations are indicators, not proof of fraud or criminal activity. Always verify information independently.