How BarryGuard Works
21 automated on-chain checks across contract security, market structure, and behavioral signals. Every check is weighted, every score is transparent.
Overview
When you submit a token address to BarryGuard, the scoring engine fetches public on-chain data from the Solana blockchain and runs 21 individual checks. Each check produces a score from 0 (maximum risk) to 100 (no risk detected). The checks are divided into three categories:
- Contract Security (6 checks): Analyzes the token's smart contract permissions and liquidity setup.
- Market Structure (7 checks): Evaluates the health of the token's trading market and holder distribution.
- Behavioral Signals (8 checks): Examines the actions and history of the token's creator and early participants.
Each check has a weight reflecting how important that signal is for rug pull detection. The total risk score is a weighted average of all 21 individual scores, clamped to a range of 1-99. No token ever receives a perfect 0 or 100 — because no on-chain analysis can be 100% certain.
Scoring and Risk Levels
The weighted average of all 21 checks produces a raw score. This score maps to five risk levels:
| Risk Level | Score Range | Interpretation |
|---|---|---|
| Danger | 1 – 29 | Critical red flags detected. Likely a rug pull or honeypot. |
| High | 30 – 54 | Significant risk signals. Proceed with extreme caution. |
| Caution | 55 – 74 | Some concerning signals. Research further before trading. |
| Moderate | 75 – 89 | Minor risk indicators. Generally acceptable but stay alert. |
| Low | 90 – 99 | No significant red flags detected. Lower risk profile. |
In addition to the weighted average, certain critical findings trigger automatic overrides. For example: if the honeypot simulation detects that a token cannot be sold, the score is forced to 1 (Danger) regardless of other checks. If the developer has a 70%+ rug rate across previous tokens, the score is forced to 1 (Danger). If the creator dumped within 30 minutes of launch, the score is capped at 5.
Three subscores are also computed — one for each category (Contract, Market Structure, Behavior). These help traders understand where exactly the risk comes from, even when the total score looks moderate.
Contract Security Checks (6)
These checks analyze the token's smart contract permissions and liquidity configuration. They answer the question: can the creator still manipulate this token at the contract level?
Mint Authority
Weight: 10What it checks: Whether the mint authority on the token's SPL account is still active or has been disabled (revoked).
Why it matters: If mint authority is active, the creator can mint unlimited new tokens at any time, instantly diluting all existing holders to zero. This is one of the most common rug pull mechanisms on Solana.
Pass: Mint authority is disabled (revoked). Score: 100.
Fail: Mint authority is still active. Score: 0.
Freeze Authority
Weight: 8What it checks: Whether the freeze authority on the token's SPL account is still active or has been disabled.
Why it matters: If freeze authority is active, the creator can freeze any wallet holding the token, preventing them from selling. This effectively traps holders.
Pass: Freeze authority is disabled. Score: 100.
Fail: Freeze authority is still active. Score: 0.
Liquidity Lock
Weight: 15What it checks: Whether the LP (liquidity pool) tokens are burnt, locked for a period, or completely unlocked. Checks for lock programs and lock duration.
Why it matters: If liquidity is not locked, the creator can remove all liquidity from the pool at any time, crashing the token price to zero. This is the classic rug pull. Burnt liquidity is the safest — it can never be withdrawn.
Pass: Liquidity is burnt (score: 100) or locked for 30+ days (score: 100).
Fail: Liquidity is not locked (score: 0). Short lock under 7 days (score: 25).
Update Authority
Weight: 6What it checks: Whether the Metaplex metadata update authority has been renounced or is still controlled by the creator.
Why it matters: If the creator still controls the update authority, they can modify the token's on-chain metadata — including its name, symbol, and image. This can be used for bait-and-switch schemes.
Pass: Update authority renounced to system program (score: 90) or controlled by third party (score: 50).
Fail: Creator still controls update authority (score: 20).
Honeypot Simulation
Weight: 12What it checks: Whether the token can actually be sold on-chain by simulating a sell via Jupiter's quote API. Measures the price impact of a test sell.
Why it matters: A honeypot is a token that can be bought but not sold — or can only be sold with extreme slippage. This check detects tokens where selling is impossible or economically unviable.
Pass: Normal sell conditions with less than 5% price impact (score: 85).
Fail: Token cannot be sold at all (score: 0). Price impact over 50% (score: 10).
LP Creator Match
Weight: 8What it checks: Whether the token creator is also the top holder of LP tokens, and whether those LP tokens are locked or burnt.
Why it matters: If the creator holds the LP tokens and they are not locked, they can remove liquidity at any time. Additional penalty applies if the top token holder and top LP holder are the same address.
Pass: Creator is not the top LP holder (score: 85), or LP tokens are burnt (score: 75).
Fail: Creator is the top LP holder and LP is not locked or burnt (score: 0).
Market Structure Checks (7)
These checks evaluate the health and maturity of the token's trading market. They answer the question: is this a real market, or an artificial setup designed to lure buyers?
Top Holder Concentration
Weight: 10What it checks: What percentage of the total token supply is held by the largest single wallet. Also calculates a Gini coefficient across the top 20 holders to measure overall distribution inequality.
Why it matters: If a single wallet holds a large share of the supply, they can crash the price by selling. Extreme concentration (over 35%) means one entity controls the token.
Pass: Top holder owns less than 5% of supply (score: 100). Under 10% (score: 85).
Fail: Top holder owns over 35% (score: 0). Between 20-35% (score: 35).
Holder Count
Weight: 5What it checks: The total number of unique wallets holding the token. Also checks the growth rate for tokens older than 24 hours.
Why it matters: A very low holder count means the token has not achieved meaningful distribution. Most rug pulls happen with fewer than 50 holders. A stagnant holder count on an older token is also a warning sign.
Pass: Over 500 holders (score: 100). 200-500 holders (score: 80).
Fail: Fewer than 10 holders (score: 0). 10-50 holders (score: 25).
Token Age
Weight: 5What it checks: How long ago the token was created on-chain, measured from the first mint transaction.
Why it matters: Brand-new tokens (minutes old) carry the highest risk because there is no track record. Most rug pulls happen within the first few hours of a token's life.
Pass: Over 7 days old (score: 100). 1-7 days (score: 60-80).
Fail: Less than 1 hour old (score: 20). Less than 6 hours (score: 40).
Liquidity Depth
Weight: 8What it checks: The actual USD value of the liquidity pool, estimated from Jupiter sell quote data.
Why it matters: Shallow liquidity means any moderately-sized trade will cause extreme price swings. A pool with less than $500 in liquidity is essentially untradeable at scale.
Pass: Pool depth over $50,000 (score: 90). $5,000-$50,000 (score: 60).
Fail: Pool depth under $500 (score: 0). $500-$5,000 (score: 30).
Bonding Curve Status
Weight: 8What it checks: Whether the token has graduated from its bonding curve (e.g., pump.fun) to a DEX like Raydium, and whether liquidity was secured upon graduation.
Why it matters: Tokens still on their bonding curve have not yet entered the open market. Graduated tokens with locked or burnt liquidity are safer. Tokens stuck on the bonding curve for over 24 hours may be abandoned.
Pass: Graduated to Raydium with liquidity locked or burnt (score: 90).
Fail: Still on bonding curve for 24+ hours (score: 15). Graduated but liquidity not locked (score: 55).
Price Impact
Weight: 10What it checks: The token's market capitalization relative to its age. Uses Jupiter-derived market cap data combined with on-chain creation time.
Why it matters: A token that is more than 24 hours old with a market cap under $1,000 has likely already been rugged. Combining market cap with age separates genuinely new tokens from abandoned ones.
Pass: Market cap over $500,000 (score: 85). $25,000-$500,000 (score: 70).
Fail: Under $1,000 market cap and over 24h old (score: 0, likely rugged). Under $5,000 and over 24h (score: 20).
Liquidity Ratio
Weight: 6What it checks: The ratio of pool liquidity (USD) to market capitalization (USD).
Why it matters: A low liquidity-to-market-cap ratio means the token is hard to trade at scale without slippage. A ratio below 1% indicates critically low liquidity relative to the token's supposed value.
Pass: Ratio above 15% (score: 85). 5-15% (score: 60).
Fail: Ratio below 1% (score: 10). 1-5% (score: 35).
Behavioral Signal Checks (8)
These checks look at the actions and history of the people behind the token. They answer the question: are the creator and early participants acting in good faith?
Developer History
Weight: 18What it checks: The creator wallet's track record — how many tokens they have previously created and how many of those are suspected rug pulls.
Why it matters: This is the highest-weighted check (weight: 18) because past behavior is the strongest predictor of future behavior. A creator with a 70%+ rug rate across previous tokens is almost certainly running a rug pull again. This check alone can force the score to Danger.
Pass: Clean track record with zero suspected rugs (score: 85).
Fail: 70%+ rug rate (score: 5, forces Danger). 40-70% rug rate (score: 30).
Insider Network
Weight: 10What it checks: Whether multiple top holders share the same funding source (fee payer) in their recent transactions, indicating they are controlled by the same entity.
Why it matters: Coordinated wallets holding a large share of the supply can dump simultaneously. If 5+ top holders share a funding source, it is a strong signal of insider manipulation.
Pass: No shared funding sources among top holders (score: 85).
Fail: 5+ holders share the same funding source (score: 0). 3-4 holders (score: 20).
Bundle Detection
Weight: 10What it checks: Whether early transactions used Jito bundles, and specifically whether the creator bundled the mint transaction with early buys.
Why it matters: Jito bundling allows creators and insiders to front-run other buyers by packing the mint and their own buys into a single atomic transaction. Creator-bundled tokens are a strong indicator of premeditated manipulation.
Pass: No Jito bundle usage in early transactions (score: 80).
Fail: Creator bundled mint with early buys (score: 0). Non-creator bundles detected (score: 30).
Early Dump
Weight: 12What it checks: Whether the token creator sold a significant portion of their holdings shortly after launch, and how quickly they sold.
Why it matters: A creator who dumps within minutes of launch is running a classic rug pull. Timing matters: dumping within 30 minutes is far worse than selling after several hours. This check can force the score to Danger if the dump happens within 30 minutes.
Pass: No early creator selling detected (score: 80).
Fail: Creator dumped within 30 minutes (score: 5, forces Danger). Within 1 hour (score: 15).
Sniper/Bot Dominance
Weight: 7What it checks: What percentage of early buy transactions matched bot timing patterns — indicating automated sniping rather than organic buying.
Why it matters: High bot dominance means the token's early market was captured by automated trading bots, not real users. These bots typically sell quickly, causing price crashes for later buyers.
Pass: No suspicious early buy concentration (score: 100). Under 20% bot share (score: 75).
Fail: Over 60% of early buys matched bot patterns (score: 0). 40-60% (score: 30).
Metadata Legitimacy
Weight: 4What it checks: Whether the token's on-chain metadata (name, symbol, image, decimals, supply) shows signs of being a scam. Checks for impersonation of major tokens, scam keywords, missing metadata, and anomalous decimal/supply values.
Why it matters: Scam tokens often impersonate well-known tokens (naming themselves “Bitcoin” or “Ethereum”), include keywords like “airdrop” or “free”, or lack basic metadata like an image.
Pass: Metadata appears legitimate with no red flags (score: 80).
Fail: Name impersonates a major token (minus 40 points). Contains scam keywords (minus 30 points). Multiple red flags can reduce score to 0.
Creator Wallet Age
Weight: 6What it checks: How old the creator's wallet is, measured from its first-ever transaction on the Solana blockchain.
Why it matters: Most rug pull creators use freshly created wallets that have no history. A wallet that is only minutes or hours old is far more suspicious than one with months of transaction history.
Pass: Wallet over 30 days old (score: 80). 7-30 days (score: 55).
Fail: Wallet created minutes ago (score: 5). Less than 1 day old (score: 15).
Creator Retention
Weight: 8What it checks: What percentage of the total token supply the creator wallet still holds, based on the top-20 holder list.
Why it matters: A creator who holds over 30% of the supply can crash the price at any time by selling. Low or zero retention by the creator is generally safer, as it means their ability to dump is limited.
Pass: Creator holds less than 1% or is not in the top 20 holders (score: 80).
Fail: Creator holds over 30% of supply (score: 15). 10-30% (score: 35).
Critical Override Rules
Some findings are so severe that the weighted average alone is not sufficient. The scoring engine applies automatic overrides in the following cases:
- Honeypot detected (cannot sell): Score forced to 1, risk level forced to Danger. If a token cannot be sold, it is worthless regardless of all other checks.
- Mint authority active + whale over 50%: Score forced to 1, risk level forced to Danger. The creator can mint infinite tokens and one wallet controls the market.
- Extreme holder concentration (over 35%): Risk level forced to Danger. Score capped at 29.
- Serial rugger (70%+ rug rate): Score forced to 1, risk level forced to Danger. The creator's track record speaks for itself.
- Creator dumped within 30 minutes: Score capped at 5, risk level forced to Danger. Classic rug pattern.
- Mixed developer history (score 6-10): Risk level raised to at least High if it would otherwise be Caution or lower.
- Unlocked liquidity: Risk level raised to at least Caution if it would otherwise be Moderate or Low.
Confidence Level
Every analysis also includes a confidence level: High, Medium, or Low. This reflects how many of the 21 checks were able to execute successfully:
- High: All 21 checks executed without errors.
- Medium: Some checks failed (data unavailable), but the majority succeeded.
- Low: More than half of the checks failed. The score should be treated with caution.
When a check cannot retrieve the data it needs (for example, if an RPC provider times out), it returns a neutral score of 40 rather than guessing. This means the overall score is slightly deflated rather than misleadingly optimistic.
Limitations
BarryGuard is a risk analysis tool, not a guarantee. Important limitations to understand:
- A low-risk score does not mean a token is a good investment. It means our checks did not find common rug pull signals. The token could still lose value for market reasons.
- A high-risk score does not mean a token will definitely rug. It means we detected signals that are commonly associated with scams.
- BarryGuard analyzes on-chain data only. Off-chain factors like social media sentiment, team credibility, or project roadmaps are not evaluated.
- Sophisticated scammers may find ways to pass some checks (for example, by locking liquidity for a short period, then rugging after unlock). Multiple passing checks together are more meaningful than any single one.
- Scores are based on the state of the blockchain at the time of analysis. The risk profile of a token can change at any time (for example, if liquidity is unlocked after a lock expires).
Always do your own research. BarryGuard is one tool in your toolkit — not the entire toolkit.
Try It Now
Paste any Solana token address into our checker and see the full 21-check analysis in seconds.
Check a TokenCopyright 2026 BarryGuard. All rights reserved.