Parallel Agent execution

[adam-paterson]

Just wanted to highlight that the orchestration and specifications produced by spec-kit are perfect for agent parallelisation!

[schlich]

@adam-paterson any recommended tools/methods for monitoring parallel execution, e.g. for identifying bottlenecks (as well as just making sure it's actually working in parallel)?

It seems like it's lacking a built-in visualization layer (which is what i'm asking about) but it would be absolutely incredible to layer spec-kit on top of something like ipyparallel or bytewax or pydantic-ai graphs

[adam-paterson]

I don't at the moment, unfortunately, but I would imagine it's out of scope for this project in particular, as this is a prompt organisation library. The actual tools executing the prompts would need the visualisation layer built into them.

Using https://docs.anthropic.com/en/docs/claude-code/hooks or https://opencode.ai/docs/plugins/, you should be able to add a small wrapper around one of those libraries.

Interesting thought though! I might look into it myself and share my findings.

[adam-paterson]

It's a little primitive, but I do monitor processes and can see increases in pids for opencode

[schlich]

yeah of course, would definitely be out of scope for the repo but moreso just asking for your personal recommendations (in case that wasn't clear, and definitely do appreciate your answer🙂)

Seems like the perfect opportunity to test out spec-kit and vibe code a solution (common refrain lately!) 😉