KERMIT

The mechanical safety checks are clean (LP locked, mint revoked, no tax, not a honeypot), but the behavioral picture is a token that has already gone quiet: 23 unique traders in 24h with sells outnumbering buys 23:11 against a ~$22k market cap means the exit is already underway and there is essentially no organic demand left. That matters far more here because RugCheck flags the creator as having a history of rugged tokens \u2014 a locked LP does not protect holders from a slow drain by a deployer who has done this before, and the flag is not reflected in the generous 80 score. The source disagreements on holder count and concentration are severe enough that neither the 'healthy 11k holders' nor the 'comfortable 16% top-10' reading can be trusted; the conservative reading (2.2k holders, ~59% in ten wallets) is far more consistent with the tiny trader count, since 11k real holders producing 23 daily traders is implausible.

• ⚠ Holder count differs by ~5x: RugCheck reports 11,426 while Birdeye (2,212) and GMGN (2,195) closely agree \u2014 the RugCheck figure is likely counting all-time or dust accounts and should be discounted
• ⚠ Top-10 concentration reported as 59.1% by RugCheck vs 16.1% by GMGN; the two cannot both be right, and the divergence likely stems from different treatment of LP/pool and burn addresses
• ⚠ RugCheck assigns a favorable score of 80 while simultaneously raising a 'danger'-level flag for creator history of rugged tokens \u2014 the score does not reflect its own worst finding
• ⚠ A claimed 11k+ holder base is inconsistent with only 23 unique traders in 24 hours, indicating the high holder figure does not represent active participants

AI reasoning over the combined data above — catches contradictions between sources that the flag-based score can't see on its own.