I don't mean to be negative, but I've been doing this already with zero extra tooling. Agents can easily use kubectl or any CLI tool (which nearly all cloud platforms offer), or use an MCP server to get running logs, and it doesn't require you to ship your production logs to a third party (which is a complete no-go for any compliance scenario btw).
fair, you’re right that agents can already pull logs via kubectl/CLIs. I wanted to build a loop tho -- capturing output, adding targeted logs when needed, rerunning, and cleaning it up end-to-end instead of stitching it together manually (in every new project)
debugy is a skill that lets your coding agent read & write logs to debug faster. during development, logs are saved locally to .debugy/session.ndjson, and the agent reads from there. for production, you can store logs in our db, and the agent can fetch them using a token.
I don't mean to be negative, but I've been doing this already with zero extra tooling. Agents can easily use kubectl or any CLI tool (which nearly all cloud platforms offer), or use an MCP server to get running logs, and it doesn't require you to ship your production logs to a third party (which is a complete no-go for any compliance scenario btw).
fair, you’re right that agents can already pull logs via kubectl/CLIs. I wanted to build a loop tho -- capturing output, adding targeted logs when needed, rerunning, and cleaning it up end-to-end instead of stitching it together manually (in every new project)
debugy is a skill that lets your coding agent read & write logs to debug faster. during development, logs are saved locally to .debugy/session.ndjson, and the agent reads from there. for production, you can store logs in our db, and the agent can fetch them using a token.