We Scanned 5 Mid-Cap DeFi Protocols. The Same Pattern Showed Up Every Time.

Here are the results from the first real batch of Token Health Scans.

First Cohort Scorecard: 5 DeFi protocols with overall scores, status, and worst dimension

Five protocols. Scores ranging from 35 to 74.

The contracts weren't broken. Something else was.

Data note: All scores based on public data as of April 30, 2026, using CoinGecko, tokenomist.ai, and DefiLlama. Three weeks into running Token Health Scan against live protocols, this is what the first cohort is showing.

The Five Protocols and Their Scores

A: Across 5 protocols, scores ranged from 35 to 74 out of 100. The protocols that scored below 60 were flagged for Tokenomics or Liquidity failures. In the one protocol with a full dimension breakdown ($MAV), Security scored 71/100 while Tokenomics scored 14/100. The contract was clean. The distribution was broken.

The full dimension breakdown for $MAV (the lowest score in this batch):

$UNI at 74/100 represents what the healthy end of this cohort looks like. No near-term unlock cliffs, reasonable holder distribution, no honeypot flags, active GitHub. Not exciting. That is the point.

For a full breakdown of how each dimension is scored, see How Token Health Scan Scores a Protocol.

The interesting data is at the bottom of the table.

Pattern 1: Security Passes. Tokenomics Breaks.

A: In this first cohort, the protocols that scored below 60 were flagged for Tokenomics or Liquidity failures, not Security. The clearest example is $MAV: Security scored 71/100 while Tokenomics scored 14/100. The contract wasn't broken. The distribution design was.

Start with $MAV. Security scored 71/100. That is a passing grade. GoPlus returned no honeypot flag. No active mint risk. No ownership backdoor.

Then look at Tokenomics: 14/100.

Here is what drove that score. More than 90% of $MAV supply sits in large wallets. The unlock schedule runs to 2060. At the time of the scan, 46.44% of the total 2 billion token supply had already been unlocked, with Advisor wallets scheduled to receive another 4.15% on July 1, 2026. Market cap was $14M against a $30M fully diluted valuation. The Tokenomics dimension scores holder concentration via Gini coefficient, upcoming unlocks, and top-wallet concentration.

The contract is not broken. Every sell pressure signal the distribution design can produce is already present.

A tool that only checks security, like Token Sniffer, would show $MAV as clean. One dimension gives a partial picture.

$ENA shows the same pattern at a different scale. The Core Contributor cliff unlock happened on May 2, 2026. TVL at the time of the unlock was $4.44B, down 69% from the $14.3B peak. The unlock event landed into a TVL environment already under significant stress. That is the difference between a cliff unlock at protocol strength versus a cliff unlock into a declining base. The overall score: 44/100.

Neither contract was broken. Both tokenomics structures were.

Pattern 2: Thin Liquidity Is the Amplifier

A: Pool depth relative to market cap determines how easily holders can exit without crashing the price. When pool depth is thin, any negative catalyst (an unlock, a drawdown, a whale exit) produces outsized price damage. $MAV's pool depth at scan time meant a $30,000 sell order moved the price 2-3%.

$DYDX scored 58/100, flagged for thin liquidity. $DYDX had seen a +54.23% price recovery over the prior 30 days. Price recovery looked constructive. Liquidity health told a different story. 83.5% of supply was already circulating, and pool depth relative to that circulating supply remained thin. The price chart and the structural health score pointed in different directions.

This is the core gap in single-metric analysis. A chart reading a +54% month and calling a token healthy misses whether there is enough depth in the market for that recovery to hold.

The clearest historical example of this gap is Iron Finance.

On June 16, 2021, Iron Finance collapsed. $2B in losses. TITAN dropped from $65 to near zero within hours. The collapse mechanism was a bank run: large liquidity providers removed IRON/USDC pool depth at scale. On-chain data confirmed retail accounts were net buyers as whale wallets exited.

The signals were visible before the cascade:

• TITAN gained 600%+ in the 7 days before collapse
• Whale LPs were reducing IRON/USDC positions
• Pool depth was thinning while social volume was climbing

Every one of these is a standard token health dimension. No single tool was watching all of them simultaneously. No one had a verdict before the cascade started.

The Iron Finance post-mortem described the event as "the world's first large-scale crypto bank run." Bank runs have early signals. The signals were on-chain. The interpretation layer was missing.

Pattern 3: The Healthy Protocol Looks Boring

A: $UNI scored 74/100 in this cohort. The profile: no near-term unlock cliffs, no honeypot flag, no concentrated whale LP positions, active GitHub, reasonable holder distribution. There is nothing dramatic in that list. That is what structural health looks like.

$UNI produces no alarming scan results. No parabolic price gains heading into an unlock. No whale concentration dominating the float. No stalled developer activity.

$PENDLE scored 66/100. Tokenomics were the cleanest in the batch, with fully linear vesting and no near-term unlock events confirmed at scan time. The flag: community lag. Not a crisis. A signal that needs monitoring.

The $PENDLE finding matters for a specific reason. A protocol can have the cleanest tokenomics design in a cohort and still have a health score below 70 if one dimension is structurally lagging. Multi-dimensional scoring produces different conclusions than single-dimension checks.

If you're building a protocol and your token health score is "boring" (distributed holder base, locked LP, active development, consistent community engagement), that is not a failure of the scoring model. That is the target.

What the Five-Tab Workflow Misses

A: The data for every finding in this batch existed in public sources before the scan. What the 5-tab manual workflow can't produce is a unified verdict. The tokenomics breakdown in $MAV was visible in holder data. The liquidity risk in $DYDX was in pool depth numbers. The cliff unlock in $ENA was documented on tokenomist.ai. Individually, each data point is noise. Combined into a single score across five dimensions, they form a diagnosis.

The standard workflow for token health analysis: Dune for on-chain queries, Nansen or Etherscan for wallet analysis, LunarCrush for social sentiment, DefiLlama for TVL, CoinGecko for price context. Each tool does one job well.

None of them produce a verdict.

After 90 minutes pulling data from those five sources, the typical output is a vague conclusion: "seems risky," "holder distribution looks concentrated," "volume is declining." That is not actionable. A protocol team needs a verdict, not a list of observations.

The Iron Finance lesson is not that the data was unavailable. Every signal was in public data before the collapse. The lesson is that fragmented data, seen one tool at a time, did not produce the compound picture needed to act.

THS collapses five dimensions into one score and a ranked remediation list. The score is the verdict. The list tells you what to fix first. See the full scoring methodology for how each dimension is weighted.

What Comes Next

A: Three patterns from this first cohort are now in the working library: Security-Passes-Tokenomics-Breaks, Thin-Float-Unlock-Ceiling, and Community-Lag-In-Otherwise-Healthy-Protocols. More scans will test whether these are consistent across the mid-cap DeFi tier or specific to this cohort.

The Day 60 target is 200 scans. At that volume, the pattern library moves from "emerging" to statistically testable.

The finding that holds up so far: the most dangerous token health states are not the ones where security is broken. They are the ones where security is fine and the distribution design is structurally flawed. The contract passes every check. The tokenomics structure produces predictable long-term sell pressure that the team didn't see, or didn't prioritize, at launch.

That is the problem Token Health Scan was built to surface.

If you issued a token in the last 24 months and you're not running regular health checks, you're operating without a diagnostic. The scan is free. It takes 60 seconds. It tells you which of the five dimensions is your current problem and what to do about it.

Run a free scan at tokenhealthscan.com.

FAQ

1. How many protocols has Token Health Scan analyzed?

The first published cohort covers 5 mid-cap DeFi protocols scanned in April 2026: $UNI, $PENDLE, $DYDX, $ENA, and $MAV. Additional scans are running on a rolling basis. Anyone can run a free scan at tokenhealthscan.com. No account required.

2. What is the most common failure mode in mid-cap DeFi token health scans?

From this first cohort, the protocols that scored below 60 were flagged for Tokenomics or Liquidity failures. The clearest example is $MAV, which scored 71/100 on Security and 14/100 on Tokenomics. The contract was clean. The distribution design was broken. This pattern is also consistent with Iron Finance, where the contract was not the failure point.

3. Can Token Health Scan predict when a token will fail?

No. THS identifies structural risk signals that are statistically consistent with known failure patterns. It does not produce price predictions or timing forecasts. A Tokenomics score of 14/100 does not tell you when $MAV will decline. It tells you the distribution structure produces consistent long-term sell pressure. What you do with that information is up to you.

4. What does a healthy token health score look like for a DeFi protocol?

$UNI at 74/100 is the closest example in this cohort. Characteristics: no near-term unlock events, reasonable holder distribution (no single wallet dominating the float), active GitHub with multiple contributors, and no security flags from GoPlus or Webacy. A score above 70 across all five dimensions indicates no critical flags are active. It is not a guarantee of future performance.

5. Is Token Health Scan for investors or protocol teams?

Both, but the tool is built for protocol teams first. Investors check one question: "Should I buy this?" Protocol founders and growth operators check a different question: "What is broken in my token, and what do I fix first?" THS is built to answer the second question. The scan output includes a prioritized remediation checklist, not a buy/sell recommendation.

References

• Iron Finance official post-mortem. June 17, 2021.
• Federal Reserve FEDS Notes: "Runs on Algorithmic Stablecoins." June 2022.
• CoinDesk post-mortem reporting. June 2021.
• CoinGecko: $MAV, $UNI, $PENDLE, $DYDX, $ENA market data. April 30, 2026.
• tokenomist.ai: $MAV, $ENA, $PENDLE unlock schedule data. April 30, 2026.
• DefiLlama: $ENA TVL data. April 30, 2026.