All blog posts
Verified Lambda Managed Instances Rust support. run_concurrent enables 8 parallel requests with 2.9ms init — effectively eliminating cold starts. Compared with standard Lambda.
Verified Lambda's GA Rust support with cargo-lambda. Cold start at 29ms, warm execution at 1.2ms — 90x faster than Python with 2.6x better memory efficiency.
Hands-on verification of AgentCore Memory Record Streaming. Covers Kinesis push delivery, FULL_CONTENT vs METADATA_ONLY behavior, and async extraction events with real test data.
Normalize metrics from multiple coding agents into a common schema and visualize them in a unified dashboard. Separating comparable from incomparable metrics is the key to avoiding misleading cross-agent comparisons.
Verifying AgentCore Runtime's new stateful MCP features. Tested elicitation (server-initiated input), sampling (LLM generation requests), and progress notifications both locally and on AgentCore.
Converted CLAUDE.md to Kiro steering and skills to Agent Skills. The key discovery: Kiro's skills follow the agentskills.io standard, so Claude Code skills can be copied directly with zero format changes.
Verified Step, Wait, Callback, and Parallel patterns via AWS CLI. Sharing checkpoint-replay behavior, gotchas, and when to choose Durable Functions over Step Functions.
A plan to roll out individually proven Claude Code meta-skills to a 20-person team. The critical design decision: separating force-distributed core skills from team-specific customizations.
Implementing 'meta-skills' to automate CLAUDE.md optimization and skill creation. Building a continuous improvement process with analyze-claude-md, optimize-claude-md, and create-skill
I cut a 355-line CLAUDE.md down to 59 lines by extracting detailed procedures into Claude Code's skills system. The key insight: separate always-loaded rules from on-demand procedures to dramatically improve response accuracy.