2 points | by warmcat 5 hours ago ago
4 comments
One thing I've been running into when working with LLM tooling is that a lot of the friction happens even before testing — just preparing context for the model.
/? mcp test: https://hn.algolia.com/?q=mcp+test :
MCP Playground, MCPSpec, MCPjam, MCP Inspector, mcp-record, mcpbr, agent-vcr
mcpbr > Supported Benchmarks : https://github.com/supermodeltools/mcpbr#supported-benchmark... :
> MCPToolBench++, ToolBench, AgentBench, WebArena, TerminalBench, InterCode; SWE-bench,
This is a good list. Are you finding these can be integrated into your CI/CD workflows?
Oh TBH I've hardly worked with MCP.
Do A2A and AP2 have better auth and authz?
There seem to be a lot of solutions for minimizing token use with MCP?
/? mcp tokens https://hn.algolia.com/?q=mcp+tokens
One thing I've been running into when working with LLM tooling is that a lot of the friction happens even before testing — just preparing context for the model.
/? mcp test: https://hn.algolia.com/?q=mcp+test :
MCP Playground, MCPSpec, MCPjam, MCP Inspector, mcp-record, mcpbr, agent-vcr
mcpbr > Supported Benchmarks : https://github.com/supermodeltools/mcpbr#supported-benchmark... :
> MCPToolBench++, ToolBench, AgentBench, WebArena, TerminalBench, InterCode; SWE-bench,
This is a good list. Are you finding these can be integrated into your CI/CD workflows?
Oh TBH I've hardly worked with MCP.
Do A2A and AP2 have better auth and authz?
There seem to be a lot of solutions for minimizing token use with MCP?
/? mcp tokens https://hn.algolia.com/?q=mcp+tokens