On-Chain Token Filters
Local, deterministic checks derived from token identity and authority signals. This stage detects suspicious symbols, known malicious authorities, dangerous authority combinations, and elevated combined risk before external market filters run.
Signals
Symbol and authority data
Scams
Known malicious authorities
Pattern
Immutable metadata plus freeze authority
Score
Combined risk threshold from 0 to 100
What the On-Chain tab is for
Use On-Chain filters as a fast security layer independent of DexScreener, GeckoTerminal, or RugCheck thresholds. These checks focus on properties that can reveal scam patterns early and produce specific reasons that remain inspectable in Status, Analytics, and Explorer.
Reading the interface
Each area answers a different operational question. Use the descriptions below before changing a filter.
Symbol analysis
Rejects malformed or suspicious token symbols commonly associated with low-quality launches.
Reject Numeric Symbols
Filters symbols made only of digits, such as 00 or 123.
Reject Empty Symbols
Filters missing or whitespace-only symbols that prevent reliable identification.
Reject Single-Char Symbols
Blocks suspicious one-character symbols when that pattern does not fit your strategy.
Authority analysis
Looks for authority ownership patterns associated with malicious control.
Reject Known Scam Authorities
Blocks tokens connected to authority addresses already identified as malicious.
Reject Immutable + Freeze
Rejects the suspicious combination of immutable metadata with an active freeze authority.
Combined risk scoring
Aggregates multiple on-chain signals into a single deterministic risk value.
Enable Combined Risk Score
Turns the weighted on-chain risk decision on or off.
Max Combined Risk Score
Rejects a token when its score reaches or exceeds the configured 0–100 threshold.
Relationship to RugCheck
On-Chain and RugCheck overlap in security intent but are separate filtering stages.
Local deterministic checks
On-Chain evaluates ScreenerBot’s own symbol, authority, and combined-risk logic.
External report checks
RugCheck contributes its report, holder, insider, LP, creator, risk-level, and transfer-fee signals.
Recommended workflow
- 1
Keep the clear identity checks enabled unless your strategy intentionally accepts unusual symbols.
- 2
Treat known scam authorities and immutable-plus-freeze as security controls, not opportunity filters.
- 3
If changing the combined score, inspect rejected examples near the current boundary in Explorer.
- 4
Compare On-Chain and RugCheck rejection pressure in Analytics rather than assuming they are duplicates.
Practical guidance
- A risk score is an aggregation of signals, not a probability of loss.
- Authority state can matter even when a token has strong volume and liquidity.
- Use exact Explorer samples to understand why a rule fires before weakening it.
Disabling authority-based checks can admit tokens whose creators retain dangerous control. More passed tokens does not mean better candidates.