LNDFlagship
Lending & Money Markets
Tested on liquidation cascades, oracle attacks, and bad-debt exposure.
Full scenario setThe test program · Phase I
Four classes. Five scenarios. No exceptions.
Every protocol in a class faces the same crashes. The results are comparable because the tests are identical.
DEX
DEX & AMMs
Tested on price manipulation, pool drains, and LP exit under stress.
Stress-testedVLT
Vaults & Yield
Tested on strategy failure, admin key risk, and withdrawal queues.
Key-risk testedBRG
Bridges & Cross-chain
Tested on validator compromise, message forgery, and frozen exits.
High-impact classRecently tested
The ratings board.
Nine of the highest-rated protocols. Open any card to read the full Crash Report.
The Crash Report · The rating
Five stars. Earned in the lab, never sold.
Every protocol gets a Crash Report: four area scores, five scenario results, and one star rating set by its weakest area. Member Reports sit beside it, never inside it.
The weakest link rule
A protocol is only as safe as its weakest area. Stars follow the lowest score, not the average.
Crash Report · CR-0001Valid
0/100 overall
Read the full Crash Report
How every report is built
Three instruments. One Test Director's signature.
Each Crash Report is run with a combined AI stack, then reviewed and signed by a human Test Director before a rating is published.
Signed by a Test DirectorNo rating is published unsigned.
TEST ENGINEER
Claude Fable 5
Reads contracts, docs, and on-chain history, runs each scenario against the Test Protocol, and drafts the scored report.
CRASH IMAGERY
Nano Banana
Produces each report's visual record: impact plates, protocol diagrams, and before/after frames.
INCIDENT DESK
Grok
Watches X and community channels live and flags exploits, depegs, and upgrades that trigger a re-test.
Run by AISigned by a humanNo rating is published unsigned
Same crash, every protocol
Every result is reproducible, and public.
In a market that runs on promises, the only rating worth trusting is one anyone can re-run.
40%Depositor Protection
25%Market Stress
15%Exit & Liquidity
20%Safety Assist
Weights set the overall score from 0 to 100. The stars ignore the weights and follow the weakest area.
Five scenarios. Fork execution is rolling out report by report.
Critical failure
If Frontal Drain or Pole Test drains funds in simulation, the rating is capped at one star, whatever the other scores.
Valid for six months
After six months the status turns Expired and the stars go grey until the protocol is re-tested.
Automatic re-test
A contract upgrade, a deployment to a new chain, or an incident sends the protocol back to the lab.
Re-runnable proof
Executed reports list their fork block and loss figures, with a script anyone can re-run.
Next class · Privacy (PRV)
Privacy needs a crash test too.
Encrypted balances hide amounts from everyone, including the people who should be checking that the money is still there. Privacy is the fifth class in the test programme. Same five scenarios. Same weakest-area rule. No stars until a report is signed.
Planned series
A crash lab that publishes its footage.
CRASH SERIES
Six bridges, one forged message. Two held.
Planned→ POST-INCIDENTFive signatures, one breach: the Ronin bridge lesson.
Published→ COMPARATIVE TESTFive lending markets under a 50% price drop.
Planned→ WEEKLY BRIEF12 protocols tested, 1 five-star.
Planned→ INTERVIEWDesigning a protocol to pass the Pole Test.
Planned→For Bulwark Members
Front-row seats. No hand on the result.
Holding the token unlocks access and early warnings. It does not, and cannot, change a rating.
BULWARKMember
Access, never stars.
Wallet 0x·· ···· ····
Member pass · illustration
Reports first
Read full Crash Reports before they open to the public.
Full test data
Rubric breakdowns, scenario details, and raw figures.
API access
Pull ratings and area scores into your own dashboards.
Re-test alerts
A notice when a protocol you use loses a star.
Member Reports & badge
File a Member Report on protocols you have used on-chain.
Member Reports and the safety rating sit side by side, and are never merged.
When users feel safe and the lab says otherwise, that is the report we publish next.
Safety rating · set by the lab
Five scenarios, four areas, one signature.
Computed from the Test Protocol. Nothing outside the lab changes it.
Lab rating
Member Reports · not part of the rating
One member, one vote.
Each report needs proof of on-chain use of the protocol and shows next to the rating, labelled separately.
0Member Reports filed so far
ONE MEMBER · ONE VOTEREPORTS REQUIRE PROOF OF ON-CHAIN USETHE TOKEN GRANTS ACCESS, NEVER STARS
In the Lab
Request a Test.
Send the protocol and its main contract. The lab picks its queue on its own, and no fee is accepted from any protocol it tests.