01
Protocols · vaults · bridges
Smart contract systems
EVM, Move, Solana, CosmWasm, and Cairo systems with complex asset and state transitions.
Independent security research & audits
Security for code that moves value.
Taichi is a competition-tested research group auditing smart contracts, ZK systems, exchanges, wallets, and the integration boundaries where critical bugs tend to hide.
2+ researchers per reviewFix review included24–48h response
Review surface / 01System online
01Accounting
02Access
03Integrations
04Economics
Invariant map
70+
Competitive audits
14
First places
170+
H/M findings
Selected first-place results
ZKsync EraMaia DAOArbitrum BoLDArcade.xyzCoinbaseOneWorld
Where we go deep
Review the system, not just the files.
The highest-impact failures often live between components: a callback that breaks accounting, a proof assumption that is not enforced on-chain, or an API boundary that changes who can sign. That is where our reviews begin.
Discuss your attack surface02
Proofs · verifiers · circuits
ZK integrations
Verifier wiring, proof-dependent state transitions, circuit assumptions, and trusted boundaries.
03
Wallets · exchanges · APIs
Web2–Web3 products
Exchange backends, wallets, APIs, signing flows, and off-chain systems touching user assets.
04
Accounting · callbacks · oracles
Integration risk
Adapters, callbacks, wrappers, share accounting, cross-protocol assumptions, and liquidity edges.
Review model
Human judgment, made systematic.
Tools widen coverage. Researchers decide what matters. Every engagement turns a protocol model into explicit paths to challenge.
Minimum two researchers
- 01
Model the system
Map assets, actors, trust boundaries, user flows, and the invariants the protocol must preserve.
- 02
Trace dangerous paths
Follow accounting changes, callbacks, liquidation branches, oracle reads, CPI boundaries, and admin operations.
- 03
Challenge the model
Two or more researchers attack the design from separate subsystems and vulnerability classes.
- 04
Verify remediation
Re-review patched code against root causes and adjacent edge cases. Fix review is included by default.
Open research
See how we reason before you hire us.
05-part internals series
Morpho, from markets to integration risk.
A source-level journey through Morpho Blue, IRMs, oracles, MetaMorpho, and the assumptions integrators inherit.
Read the work
04-part security series
Solana attack surfaces, traced.
Account creation, Anchor initialization, stale CPI state, and Token2022 semantics.
Read the work
Incident analysis
One deposit. Two mints. $2.7M lost.
The Solv callback path reconstructed from source to exploit.
Read the work
Cryptography
Invalid curves and small subgroups.
The group theory and implementation mistakes behind key-recovery attacks.
Read the workVerifiable signal
Trust the record, not the adjectives.
Security marketing is easy to write. Public placements, reproducible research, and findings that survive judging are harder to fake.
70+
Competitive audits
14
First places
24
Top 3 finishes
170+
H/M findings
Contest record
0124 Top 3 finishes
14 first-place results across high-complexity protocol reviews.
Adversarial practice
02Remedy CTF 2025 · #7
Ongoing CTF work keeps exploit construction—not just issue classification—in the review loop.
Published reasoning
03Research you can inspect
Source walkthroughs and incident analyses expose the depth and clarity of our reasoning.
Every audit includes two or more researchers and a focused fix review by default.
No add-on feePublic portfolio data · verified Jul 2026
Start a review
Your code is nearly ready. Let's make it hard to break.
Share the repository scope, target chain, key integrations, and launch window. We typically reply with availability and next steps within 24–48 hours.
Open audit requestDirect channels
Intake open
Critical disclosure involving a protocol we reviewed? Use Telegram or email for the fastest response.