Aft, a Python toolkit to study agent behavior
Behavioral field approximation via trajectory sampling; clever framing, limited adoption signals.
Do MCP tools serialize in Claude Code? Empirical study: readOnlyHint controls parallelism, IPC overhead is ~5ms/call. Reproduces #14353.
Benchmarks readOnlyHint's impact on MCP parallelism, but archived before yielding actionable best practices.
MCP server authors, Claude Code users, researchers studying AI tool parallelism
Behavioral field approximation via trajectory sampling; clever framing, limited adoption signals.
Parallel Chrome contexts reuse your logins; 80x faster than spawning fresh Playwright instances.
Packs 359MB of scholarly lexicons into an MCP server for instant AI biblical analysis.
Reclaims wasted context tokens by auditing MCPs and skills before you type.
Running 15 concurrent agents without burning through API limits faster than CrewAI or AutoGen.
Blog post about building MCP UIs, not an actual tool you can use today.