Skip to content

Robinhood Chain·Scanned 05 Nov 2026, 19:39 UTC

PonsV2LauncherToken $KILLFLOOR

0xaf2d6cc1586539997840d4635480c2f8494dac2e

Verdict

Scam

24/100 safety

Token has no owner powers and sells at 0% tax, but the app's redeem test failed and insiders hold 31% of supply.

The KILLFLOOR token contract itself is clean: there is no owner, no mint and no tax, buying and selling work, and the main pool's liquidity is permanently locked. The rating is pulled down by the Killfloor app around it. In our test, withdrawing a deposit from an agent could not be completed, and deposits sit in unverified contracts whose owner can pause the system or change redemption rules. The launch was also captured by sniper and creator-declared wallets that still hold about 31% of supply and could sell into a small pool.

Score breakdown

  • Contract
    15
  • Token
    22
  • Social / X
    45
  • Website
    40

Higher is safer · 0 = worst · 100 = best

Yes
0%
None (no owner role)
0% / 100%

NarrativeSteady

Trend scout · Grok 4.7 · X search via twitterapi.io

46/100 heat

A separate score, not part of the safety score: hype never makes a contract safer.

  • AI agents
  • agent commerce
  • AI trading
Fit
Core: the product is this narrative
Buzz
Mixed50 posts from 39 accounts; 5% of authors have accounts under 30 days old; 16 post(s) with engagement that does not fit the follower count.
Momentum
Project mentions roughly 460 vs 667 the prior week (flat to slightly lower, rough); AI-agent talk roughly flat (~4472 vs ~4918).

Killfloor is talked about as a Robinhood Chain AI-agent city where agents trade and pay each other, including a Gloam private-payment collab. Mentions are steady to slightly lower week over week, not a dominant topic. Much of the recent chatter is list mentions, collab replies, and low-engagement tags rather than broad discussion.

Posts and accounts (5 posts)
  • @gloamtrade · 05 Nov 2026, 16:13 UTC

    “Killfloor’s AI agents now make private payments via Gloam.”

  • @Crypt0WhaleX · 29 Oct 2026, 16:55 UTC

    “$KILLFLOOR is a live 3D city where every AI agent stands as its own tower, trading real money.”

  • @winiam4444 · 29 Oct 2026, 14:39 UTC

    “1. Killfloor, $KILLFLOOR, @klll_floor Hit $250K MCAP - Now $210K MCAP Category: AI Agents”

  • @HOODdailyTK · 29 Oct 2026, 14:52 UTC

    “Spotlight Memecoin on Robinhood Ecosystem Last 24h”

  • @WalkDog1984 · 01 Nov 2026, 18:57 UTC

    “the most sniped insider traded projects on $pons are #1 $killfloor and #2 $grift”

Who pushes it

Competing for the same narrative

Lobin, Wagie, Grift

Posts, last 7 days vs the 7 before

  • project: ~460 vs 667 (flat (rough))
  • AI agents: ~4472 vs 4918 (flat (rough))
  • agent commerce: ~882 (no comparison)

~ = estimated from one page of results

Measured in code from 50 posts by 39 accounts: 5% fresh accounts (< 30 days), 0% near-copies, top 3 accounts wrote 16%.

X posts searched from 2026-10-06 to 2026-11-06 (UTC). Post links come only from the X search results.

MC at check
$50.9K
MC now
no live price
Since check
—
Liquidity
$19.8K

$KILLFLOOR · Live price chart

Live chart

The chart updates live; “MC at check” above is the market cap when this check ran.

LIVE
DexScreener · uniswap · pair 0x89a3…8230

Key findings

  • CriticalWithdrawing through the dApp failed ("Redeem agent shares")
  • HighClaim contradicted: The 5% fee from Fade bets is earmarked for future KILLFLOOR token buybacks.
  • HighDump risk: launch wallets still hold 31.3% of supply
  • HighCreator-declared wallets bought 55.6% of supply at launch
  • GoodToken is a fixed-supply Pons template with no owner powers

What would change the verdict

  • Raises safety:Agent vault and registry source is verified and a redemption succeeds on a fork, so the dApp exit test passes
  • Lowers safety:Creator-declared insider and bundle wallets sell their remaining ~31% into the ~$19.9k locked pool
  • Lowers safety:App owner calls setPaused or raises the redeem cooldown while backer funds sit in agent vaults
  • Raises safety:Fade fees are actually used to buy back and burn KILLFLOOR on-chain, as the project claims

Unlock PRO for this project

A one-time payment adds hourly re-checks for good, a 7-day watch, questions to the jury, the launch map and deeper social forensics.

Post on XView on explorer ↗

Findings

17 findings

Risks (10)

  • CriticalWithdrawing through the dApp failed ("Redeem agent shares")honeypotDeterministic

    We ran the project's own dApp on a local fork. The "Redeem agent shares" step did not work. Expected: Redeem in kind (always available) or for dollars only (up to 1% slippage). Redemption in kind needs no price and no backend, so backers can always exit. Redeem your shares after a short cooldown, in kind or for dollars only. (source: https://killfloor.org/faq, https://killfloor.org/docs, https://killfloor.org/). Observed: wallet no balance change; 0 tx (none). No successful transaction was made for this goal (timeout). | driver: timeout — goal time cap reached. Money deposited may be impossible to take out.

    Evidence

    • forkdApp test "Redeem agent shares" on a local fork (https://app.killfloor.org/floor)
      Show source
      wallet no balance change; 0 tx (none). No successful transaction was made for this goal (timeout). | driver: timeout — goal time cap reached

    Confirmed by: facts

  • HighClaim contradicted: The 5% fee from Fade bets is earmarked for future KILLFLOOR token buybacks.claims-mismatchDeterministic

    The project claims token burns / buybacks, but no tokens of this contract were ever sent to a burn address or destroyed.

    Evidence

    Confirmed by: facts

  • HighDump risk: launch wallets still hold 31.3% of supplyholdersDeterministic

    Sniper / insider / bundled launch wallets still hold about 31.3% of the supply. If they sell, the price can collapse — current dump pressure is HIGH.

    Evidence

    • on-chainheld now: snipers 30.3189%, insiders 31.32%, clusters 19.7%

    Confirmed by: facts

  • HighCreator-declared wallets bought 55.6% of supply at launchinsidersDeterministic

    32 wallet(s) linked to the team (declared by the creator at launch, exempt from the snipe tax) bought 55.6% of the supply in the first blocks — an insider allocation that real buyers did not get. They now hold 31.3%, so current dump pressure from them is HIGH.

    Evidence

    • on-chaininsiders: 32 wallets, fundedBy declared by creator at launch (snipe-tax exempt on curve; creator 0x29d86c12dc00d6257bde7ecc4160e9b79468efdd), bought at launch 55.59%, pctSupplyHeldNow 31.3189
    • on-chaintoken.insiders[0].fundedBy / pctSupplyHeldNow
      Show source
      declared by creator at launch (snipe-tax exempt on curve; creator 0x29d86c12dc00d6257bde7ecc4160e9b79468efdd) — pctSupplyHeldNow 31.3189
    • on-chaintoken.liquidity[0].liquidityUsd
      Show source
      19857

    Confirmed by: facts, claude-opus-5.5, grok-4.7, gpt-5.6-sol

  • HighLaunch sniped: 33 wallets took 64.3% of supply in 11 secondssnipersDeterministic

    The launch was captured: 33 wallets bought 64.3% of the supply within 11 seconds. A small, fast group set the early price and could sell into everyone who bought later — bad for launch fairness and a strong insider/bot signal. They now hold 30.3% of supply, so current dump pressure from them is HIGH — most of it was already sold, i.e. they exited on later buyers.

    Evidence

    • on-chainsnipers: windowBlocks=100, wallets=33, pctSupplyBought=64.3086, pctSupplyHeldNow=30.3189
    • on-chaintoken.bundles[0] (block 100350783)
      Show source
      commonFunder 0xa5a5491bca93dd4c076e4906e79e7673f4a5a142, pctSupply 38.3118
    • on-chaintoken.bundles[1] (block 100350784)
      Show source
      commonFunder 0x728f51950ff096dc03a48f5e83f45ac8bb75f500, pctSupply 17.2756
    • on-chaintoken.snipers
      Show source
      pctSupplyBought 64.3086, pctSupplyHeldNow 30.3189
    • on-chaintoken.clusters[0]
      Show source
      funder 0x728f51950ff096dc03a48f5e83f45ac8bb75f500 ... pctSupplyHeldNow 8.4751

    Confirmed by: facts, claude-opus-5.5, grok-4.7, gpt-5.6-sol

  • HighAgent deposits sit in an unverified vault with kill/trade functionsotherAI

    In the dApp test, backing an agent approved exactly 10 USDG and called requestDeposit on the agent contract 0xEd4b…f938 ('KF-3 token'). That contract now holds the 10 USDG. Its source is unverified. Its function list includes kill(), retire(), trade(address,address,address,uint256,uint256,bytes), crystallize() and claimOwed(address). Nobody can check who may call these or where trade() can send the deposited funds. The site promises 'backers can always exit' by redeeming in kind, but the redeem test produced no transaction (UI timeout). That exit path is therefore unproven, not proven broken. Backers must fully trust unverified code and its operators with their deposits.

    Evidence

    • forkdappTests 'Back agent' tx requestDeposit
      Show source
      wallet -10 USDG → 0xed4b…f938 (KF-3 token) [unverified contract — third party?]
    • forkdappTests 'Redeem agent shares'
      Show source
      could not complete "Redeem agent shares" in the UI (timeout: goal time cap reached)
    • staticrelated KF-3 token 0xEd4bEa49e2650Ae96b9d13a7cF5D70624089f938 function table
      Show source
      fn kill() [unknown (unverified)] risk=critical

    Confirmed by: claude-opus-5.5, grok-4.7 · Disputed by: gpt-5.6-sol

  • HighApp owner can pause agents and change redeem cooldown and limitspauseAI

    The app contract 0x37A0…a3d5 is unverified and owned by 0x7370…CD1E. It looks like the agent registry/factory: it has vaultOf, vaultImplementation and agentCount. It exposes setPaused(bool), setRedeemCooldown(uint64), setLimits(...), setAdapter(address,bool), setRoles(address,address,address) and setAssetClass. The owner could pause the system, lengthen the redeem cooldown, or approve new trading adapters that agents can route funds through. Any of these could delay or block backers' exits. No bounds or timelock can be verified, because the source is not published. This does not affect the KILLFLOOR token itself.

    Evidence

    • staticrelated 0x37A0b894b7636D6B87B6aA4055427665cBBba3d5 function table
      Show source
      fn setPaused(bool) [unknown (unverified)] risk=high; fn setRedeemCooldown(uint64) ... risk=medium; fn setLimits(...) risk=medium

    Confirmed by: claude-opus-5.5, grok-4.7 · Disputed by: gpt-5.6-sol

  • MediumApp price oracle is fed by an owner-appointed posterotherAI

    The unverified contract 0x2615…8eEc is owned by 0x7370…CD1E. It has post(address[],uint256[]) and priceE18(address), and the owner can call setPoster, setMaxMoveBps and setMaxAge. The site's agent valuations (NAV, share prices, bet settlement) likely depend on these posted prices; that link is inferred, not verified. If so, the owner could appoint a poster and loosen the max-move limit. Posted prices could then misvalue agent shares at deposit or redemption.

    Evidence

    • staticrelated 0x26157d6B41587d6Ef36db16251383B0F14108eEc function table
      Show source
      fn setMaxMoveBps(uint16) [unknown (unverified)] risk=high; fn setPoster(address); fn post(address[],uint256[])

    Confirmed by: claude-opus-5.5, grok-4.7

  • LowIBM partnership is claimed only by the project itselfclaims-mismatchAI

    On 29 Oct 2026 the project said it had only applied to IBM Partner Plus. Six days later it announced official acceptance and access to IBM technology. A search of posts about that claim returns only Killfloor’s own announcements. IBM, IBMPartners, and IBMDeveloper are tagged but did not confirm it in the results. Treat the partnership as unverified marketing, not as proof of enterprise backing.

    Evidence

    Confirmed by: grok-4.7

  • LowFade fee described differently on X and on the websiteclaims-mismatchAI

    The X account says every Fade bet carries a 5% fee that goes to a city treasury for future token buybacks. The website says the fade fee is taken from winnings, never from stakes, and lists a separate 5% protocol fee on profits above a high-water mark. Holders cannot tell from these two sources whether a bet is charged 5% up front or only winners pay.

    Evidence

    Confirmed by: grok-4.7, gpt-5.6-sol

Positive signals (3)

  • PositiveToken is a fixed-supply Pons template with no owner powersownershipAI

    KILLFLOOR is the standard PonsV2LauncherToken: a plain OpenZeppelin ERC20 plus a voluntary burn function. It has no owner, mint, fee, blacklist, pause or proxy. The whole supply was minted once to the bonding curve in the constructor. 'deployer' is stored only for information and gives no privileges. The custom code only adds read-only metadata functions. Slither found no issues.

    Evidence

    • codePonsV2LauncherToken.sol:L21-L71
      Show source
      contract PonsV2LauncherToken is ERC20, ERC20Burnable {
          struct Socials {
              string twitter;
              string telegram;
              string discord;
              string website;
              string farcaster;
          }
      
          error ZeroAddress();
      
          address public immutable deployer;
          address public immutable launchFactory;
          address public immutable curve;
    • codeERC20.sol:L176-L204
      Show source
      function _update(address from, address to, uint256 value) internal virtual {
              if (from == address(0)) {
                  // Overflow check required: The rest of the code assumes that totalSupply never overflows
                  _totalSupply += value;
              } else {
                  uint256 fromBalance = _balances[from];
                  if (fromBalance < value) {
                      revert ERC20InsufficientBalance(from, fromBalance, value);
                  }
                  unchecked {
                      // Overflow not possible: value <= fromBalance <= totalSupply.
                      _balances[from] = fromBalance - value;
                  }
              }
    • staticstatic.issues = [] (slither 0.11.6)

    Confirmed by: claude-opus-5.5, grok-4.7, gpt-5.6-sol

  • PositiveFork test: buys and sells work at 0% token taxhoneypotAI

    A real buy and sell on a fork of the graduated Uniswap v4 pool both succeeded, with output exactly as expected. A second wallet that received tokens by transfer could also sell, and selling after one hour worked too. Token-level buy, sell and transfer tax are all 0%.

    Evidence

    • forksimulation.notes
      Show source
      second wallet (received via transfer) could sell; sell after +1h: ok

    Confirmed by: claude-opus-5.5, grok-4.7, gpt-5.6-sol

  • PositiveMain pool liquidity is permanently locked by the Pons lockerliquidityAI

    All liquidity of the graduated pool 0x89a3a698… is held by the PonsV2LaunchLocker, a permanent LP lock. The deployer holds none of it, so the main liquidity cannot be pulled. Three other small KILLFLOOR v4 pools exist with unlocked liquidity, but they hold only $20, $2 and $0.

    Evidence

    • on-chaintoken.notes: v4 position liquidity holders of 0x89a3a69840…
      Show source
      0x267444d099b10fb5ed7c3cc7b7c767adca574952 [locked: PonsV2LaunchLocker (permanent LP lock)] 100%
    • on-chaintoken.liquidity[0]
      Show source
      lpLockedPct 100, lpHeldByDeployerPct 0

    Confirmed by: claude-opus-5.5, grok-4.7, gpt-5.6-sol

Notes (4)

  • MediumMain pool hook can alter swaps; its code was not reviewedotherAI

    The main KILLFLOOR pool is attached to the hook 0xE5e7…e044 (PonsV2MemeHook). The hook has the permissions beforeInitialize, afterSwap and afterSwapReturnsDelta, so after every swap it can take a cut of, or change, the swap output. The fork simulation measured 0% token-level buy and sell tax and no reverts, so today it does not block or heavily tax trades. The hook source and its admin settings were not provided, so I cannot verify who could change its behavior or where any fee goes. The launchpad notes say 'hook fees may apply' after graduation.

    Evidence

    • on-chaintoken.notes: uniswap-v4 0x89a3a69840… hook permissions
      Show source
      v4 hook 0xE5e702641Ea86F4ae6cC3cDaeD2B886f976Be044 permissions: beforeInitialize, afterSwap, afterSwapReturnsDelta
    • forksimulation buyTaxPct/sellTaxPct
      Show source
      canBuy true, buyTaxPct 0, canSell true, sellTaxPct 0
    • on-chainlaunchpad.fees.notes
      Show source
      after graduation trades go through the Uniswap v4 pool (hook fees may apply)

    Confirmed by: claude-opus-5.5, grok-4.7, gpt-5.6-sol

  • InfoDeployer holds 1% and has not solddeployer-historyAI

    The deployer wallet 0x29D8…Efdd holds 1% and has sold 0%. No earlier deployer history was recorded. Note, however, that the deployer is part of the creator-declared insider group.

    Evidence

    • on-chaintoken.notes / deployerSoldPct
      Show source
      Deployer 0x29d86c12dc00d6257bde7ecc4160e9b79468efdd holds 1% now; deployerSoldPct 0

    Confirmed by: claude-opus-5.5, grok-4.7, gpt-5.6-sol

  • InfoAnonymous team; X account about two weeks oldsocialAI

    The public face is @klll_floor, created 21 Oct 2026, verified, with 198 followers. The site and posts name no person. The domain was created 7 Oct 2026 and the token was deployed 28 Oct 2026. That fits a new anonymous project, not a bought old account, and it is not by itself proof of a rug. Nobody is publicly accountable if the product or token claims fail.

    Evidence

    Confirmed by: grok-4.7

  • InfoProject says its app contracts are unaudited and its terms are a draftwebsiteAI

    Killfloor’s own docs say the protocol contracts have not had an external audit, and the terms page is labeled a draft pending legal review. The site also says a compromised backend cannot withdraw backer funds and that sells and redemptions stay open if the app is paused. Those are disclosures about the agent product, not a third-party check, and they do not describe owner powers on the KILLFLOOR token itself.

    Evidence

    • Show source
      The contracts have not had an external audit yet. When one is done, the full report and every fix will be linked here.
    • Show source
      The backend holds separate keys to trade, post prices and bill fuel. None of them can withdraw backer funds.
    • Show source
      Terms of use for the Killfloor website and app (draft pending legal review).

    Confirmed by: grok-4.7, gpt-5.6-sol

Score breakdown

Higher is safer. The overall is the weighted average of the assessed parts, then capped by hard facts.

  • Contract

    weight 45%
    15
    Capped · max 15

    dApp test "Redeem agent shares" failed — funds may not be withdrawable → max 15

    The KILLFLOOR token itself is a clean, ownerless Pons template that trades at 0% tax, but the score is capped because the project's dApp redeem test failed and its deposit contracts are unverified and owner-controlled.

    Why this score

    Raised the score

    • Fixed-supply Pons template, no owner, mint, fee, blacklist or pause
    • Verified source, not a proxy, Slither found no issues
    • Fork buy and sell work at 0% token tax, including transferred wallets and after 1h

    Lowered the score

    • Redeem agent shares test failed in the dApp (no transaction completed)
    • Agent deposits held in an unverified vault with kill/trade functions
    • App owner can pause, change redeem cooldown/limits and adapters
    • Owner-appointed price poster for the app oracle
    • Pool hook can alter swaps; code not reviewed
  • Token

    weight 30%
    22
    Capped · max 45

    launch wallets still hold 31.3% of supply → max 45

    The launch was captured: snipers bought 64% and creator-declared insiders 56% of supply in seconds, and they still hold about 31% against a small ~$19.9k pool.

    Why this score

    Raised the score

    • Main pool liquidity 100% permanently locked in the Pons locker
    • Deployer holds 1% and has not sold

    Lowered the score

    • 33 sniper wallets took 64.3% of supply in 11 seconds
    • 32 creator-declared wallets bought 55.6% at launch, still hold 31.3%
    • Bundled same-block buys from common funders took over half of supply
    • Claimed buybacks not seen on-chain
    • Thin liquidity (~$19.9k) relative to insider holdings
  • Social / X

    weight 15%
    45

    A new, anonymous project whose X account is two weeks old, with an unconfirmed IBM partnership and inconsistent fee descriptions, though there are no signs of bought engagement.

    Why this score

    Raised the score

    • X account is new but not a repurposed old account
    • No credible rug or honeypot accusations found

    Lowered the score

    • Team is unnamed, nobody publicly accountable
    • IBM Partner Plus acceptance only self-claimed
    • Fade fee described differently on X and the site
  • Website

    weight 10%
    40

    The site is functional and discloses that its contracts are unaudited, but its buyback claim is contradicted by the facts and its 'backers can always exit' promise was not shown to work.

    Why this score

    Raised the score

    • Detailed docs, FAQ and risk pages
    • Openly states the contracts are unaudited and the terms are a draft

    Lowered the score

    • Domain created Oct 2026 (older archive history belongs to a previous owner)
    • Buyback claim contradicted on-chain
    • Promised redemption could not be demonstrated in the dApp test