Reduce MCP tooling context usage

[rcdailey]

Each of your MCP tools consumes over 1k tokens which is unnecessary and likely due to the highly verbose AI-generated descriptions for each tool. Please optimize the token usage of your MCP tools by greatly compacting and optimizing text based descriptions. Here is a context analysis report from Claude Code that shows specific numbers:

[bgauryy]

Thanks for letting me know!
It will be reduced (and I'm in the process of making some more optimizations)

It's great feedback!

[rcdailey]

octocode is a fantastic tool; thank you for it!

[bgauryy]

@rcdailey - FYI
WIP PR - https://github.com/bgauryy/octocode-mcp/pull/128/files#diff-98ea69508c3b3ac54179e1f86ba322718fd083341272a1c95ece54816aa1a071

And by the way - its scheme complexity issues and some research tradeoff.
Descriptions are not generated from AI...it's the hardest part in this MCP :)

[rcdailey]

That's still a lot, IMO, but I understand it's a challenge to optimize. Thank you for addressing it and I'm sure you'll continuously improve it over time!

[bgauryy]

For sure!
I will optimize it more and will keep this issue opened until then (I want to get to less than 1k tokens for each tool in context).

Updated version
octocode-mcp@5.0.0-alpha.6

[bgauryy]

@rcdailey
Reduced to 4k total on 6.1.0
Closing the issue for now :)

Will work on improving it more
Thanks for the issue!

[rcdailey]

Great work, thank you! Your latest github release still shows 4.x. When will you release a new stable version? I am using @latest and not any prerelease version.