Skip to main content

Limitations

Rug Radar improves speed and clarity, but it is not omniscient.

Important Limits​

  • Onchain data can lag or be incomplete.
  • Some behavior heuristics are probabilistic, not ground truth.
  • Fast-moving tokens can change materially in minutes.
  • Partial source outages can reduce confidence.
  • Name-based token lookup can resolve to the most liquid matching Solana pair, which may not be the asset you intended.
  • Wallet analysis is heuristic-heavy and should not be treated as a verified PnL model.
  • Monitoring is snapshot-based, not continuous stream processing.

Tool-Specific Limits​

analyze_token​

  • Risk scores are deterministic but still depend on upstream freshness
  • token age comes from DEX pair creation time, not necessarily the true mint creation time
  • missing holder data or contract data lowers confidence and can distort interpretation

scan_risk​

  • This is intentionally opinionated and threshold-driven
  • it can miss nuanced situations that show up only in the full analyze_token output

analyze_wallet​

  • Wallet type labels are heuristic classifications
  • win rate is currently a rough placeholder rather than a realized PnL engine
  • average hold time is an approximation

watch_token and get_token_changes​

  • Alerts are only evaluated when get_token_changes runs
  • no background scheduler is included in this repository
  • baseline and delta logic depend on a working Postgres connection

Operational Limits​

  • No built-in auth middleware is enabled in the current server
  • The server caches upstream responses in memory only
  • A process restart clears that cache
  • Redis is used only for alert deduplication, not as a primary datastore

Best Practice​

  • Re-run checks before entry
  • Use watch + change tracking for volatile assets
  • Combine with your own thesis and risk management
  • Prefer token addresses over names when precision matters
  • Treat low-confidence outputs as incomplete, not reassuring