While we wait for the dust to settle and hope for the best there is, sharing a few pointers on hardening the admin control flow for programs:
1. Program-native multisig for all admin config ixs over and above the obvious external multisig/cold-wallet setup
2. Timelocked admin execution enforced onchain for critical changes to program config like new listings, thresholds for circuit breakers etc.
3. Real-time admin config alerts so team/community can act on time critical intel.
4. No-op buffer for emergency bricking so program upgrade multisig can swap in a single tx
Drift is one of the multiple credit venues that Asgard sources from.
Our current open interest routed through Drift is not significant.
On our end, we have disabled Drift as a credit venue on Asgard and are reaching out to affected users directly.
We're in contact with the Drift team and monitoring the situation closely.
No other Asgard operations are affected, we will be sharing updates as we learn more.
🚨 CRITICAL: Active supply chain attack on axios -- one of npm's most depended-on packages.
The latest [email protected] now pulls in [email protected], a package that did not exist before today. This is a live compromise.
This is textbook supply chain installer malware. axios has 100M+ weekly downloads. Every npm install pulling the latest version is potentially compromised right now.
Socket AI analysis confirms this is malware. plain-crypto-js is an obfuscated dropper/loader that:
• Deobfuscates embedded payloads and operational strings at runtime
• Dynamically loads fs, os, and execSync to evade static analysis
• Executes decoded shell commands
• Stages and copies payload files into OS temp and Windows ProgramData directories
• Deletes and renames artifacts post-execution to destroy forensic evidence
If you use axios, pin your version immediately and audit your lockfiles. Do not upgrade.
the issue is that we’re assuming a black box will follow a state machine.
for context: even the researchers building these neural systems don’t fully understand how they work, they’re still largely a black box.
views are my own.
p.s. claude still missed my workflow step :cry
@simpdigit right, but my line of thinking is:
can we build an orchestrator where workflow execution is deterministic like a math equation, where invalid transitions are not just disallowed, but unrepresentable?
just thinking out loud
the issue is that we’re assuming a black box will follow a state machine.
for context: even the researchers building these neural systems don’t fully understand how they work, they’re still largely a black box.
views are my own.
p.s. claude still missed my workflow step :cry
@simpdigit yes, but the next step up is that we’ll need some kind of state machine orchestrator that enforces things very strictly.
I don’t think this orchestrator can be another LLM or AI, though. It likely needs to be something else, I’m just not sure what that is yet 👀
I did look into this in some detail a while back.
In CFMMs, LVR is structural because the AMM has no external price reference. It’s always behind the market, and arbitrage is what updates the price, so latency-driven loss is inherent.
In pool-based perps like GMX or Jupiter, pricing is oracle-driven. the system is already anchored to external prices, so it doesn’t rely on arbitrage for price discovery. any LVR-like effect mainly comes from small windows of oracle or execution latency, which are bounded.
so while some leakage exists, it’s relatively minor. In practice, LP PnL is driven more by trader flow and positioning than by systematic price lag.