Futures
Access hundreds of perpetual contracts
TradFi
Gold
One platform for global traditional assets
Options
Hot
Trade European-style vanilla options
Unified Account
Maximize your capital efficiency
Demo Trading
Introduction to Futures Trading
Learn the basics of futures trading
Futures Events
Join events to earn rewards
Demo Trading
Use virtual funds to practice risk-free trading
Launch
CandyDrop
Collect candies to earn airdrops
Launchpool
Quick staking, earn potential new tokens
HODLer Airdrop
Hold GT and get massive airdrops for free
Launchpad
Be early to the next big token project
Alpha Points
Trade on-chain assets and earn airdrops
Futures Points
Earn futures points and claim airdrop rewards
Claude Code SDK's --bare flag cuts startup time by 10x
Headline
Claude Code SDK gets --bare flag for 10x faster startup in scripts and CI pipelines
Summary
Boris Cherny, who leads Claude Code at Anthropic and built the tool, tweeted about a performance fix for the Claude Agent SDK. The --bare flag skips the default scans for local config files (CLAUDE.md for project context, settings, MCPs for tool extensions) and cuts startup time by up to 10x.
This matters for non-interactive use cases - scripts, CI pipelines, batch jobs - where you’re already specifying everything explicitly through flags like --system-prompt or --mcp-config. Loading those default configs was always unnecessary overhead; the SDK just did it automatically.
Analysis
The fix addresses a real pain point for anyone running Claude agents in automated environments. When you’re spinning up agents in a loop or triggering them from CI, that startup overhead compounds fast. A 10x improvement isn’t theoretical - it’s the difference between “sluggish” and “usable” for batch processing or serverless setups.
Cherny mentioned they’re planning to make --bare the default in future versions, which suggests Anthropic is listening to how developers actually use the SDK. Most programmatic use cases don’t need automatic config discovery - that’s more of a convenience for interactive development.
The broader context here: AI tooling is still catching up to production requirements. Features get built for the demo case first, then optimized for real workloads later. This is that “later” happening.
Impact Assessment