I had deepseek explain MCP to me. Then I asked what was the point of persistent connections and it said it was pretty much hipster bullshit and that some url to post to is really enough for an llm to interact with things.
red_hare 17 minutes ago [-]
True for coding agents running SotA models where you're the human-in-the-loop approving, less true for your deployed agents running on cheap models that you don't see what's being executed.
I’m not sure how this works. A lot of that tool description is important to the Agent understanding what it can and can’t do with the specific MCP provider. You’d have to make up for that with a much longer overarching description. Especially for internal only tools that the LLM has no intrinsic context for.
thellimist 13 minutes ago [-]
I can give example.
LLM only know `linear` tool exists.
I ask "get me the comments in the last issue"
Next call LLM does is
`linear --help 2>&1 | grep -i -E "search|list.issue|get.issue")`
then
`linear list-issues --raw '{"limit": 3}' -o json 2>&1 | head -80)`
then
`linear list-comments --issue-id "abc1ceae-aaaa-bbbb-9aaa-6bef0325ebd0" 2>&1)`
So even the --help has filtering by default. Current models are pretty good
hiccuphippo 10 minutes ago [-]
Can LLMs compress those documents into smaller files that still retain the full context?
andybak 12 minutes ago [-]
Why are they using JSON in the context? I thought we'd figured out that the extra syntax was a waste of tokens?
mijoharas 18 minutes ago [-]
This sounds similar to MCPorter[0], can anyone point out the differences?
- written in go
- zero-dependency binaries
- cross-compilation built-in (works on all platforms)
- supports OAuth2 w/ PKCE, S2S, Google SA, API key, basic, bearer. Can be extended further
MCPorter
- TS
- huge dependency list
- runtime dependency on bun
- Auth supports OAuth + basic token
- Has many features like SDK, daemons (for certain MCPs), auto config discovery etc.
MCPorter is more complete tbh. Has many nice to have features for advanced use cases.
My use case is simple. Does it generate a CLI that works? Mainly oauth is the blocker since that logic needs to be custom implemented to the CLI.
speedgoose 24 minutes ago [-]
MCP has some schemas though. CLI is a bit of a mess.
But MCP today isn’t ideal. I think we need to have some catalogs where the agents can fetch more information about MCP services instead of filling the context with not relevant noise.
thellimist 18 minutes ago [-]
It's the same from functionality perspective. The schema's are converted to CLI versions of it. It's a UI change more than anything.
groby_b 6 minutes ago [-]
You are free to build tools that emit/ingest json, and provide a json schema upon request.
The point is push vs pull.
crooked-v 41 minutes ago [-]
Cheaper, but is it more effective?
I know I saw something about the Next.js devs experimenting with just dumping an entire index of doc files into AGENTS.md and it being used significantly more by Claude than any skills/tool call stuff.
thellimist 12 minutes ago [-]
personal experience, definitely yes. You can try it out with `gh` rather than `Github MCP`. You'll see the difference immediately (espicially more if you have many MCPs)
dang 13 minutes ago [-]
The article's link to clihub.sh is broken. Looks like https://clihub.org/ is the correct link? I've added that to the toptext as well.
thellimist 10 minutes ago [-]
Good catch.
I didn't release the website yet. I'll remove the link
Rendered at 21:33:47 GMT+0000 (Coordinated Universal Time) with Vercel.
But yeah, a concrete example is playwright-mcp vs playwright-cli: https://testcollab.com/blog/playwright-cli
LLM only know `linear` tool exists.
I ask "get me the comments in the last issue"
Next call LLM does is
`linear --help 2>&1 | grep -i -E "search|list.issue|get.issue")` then `linear list-issues --raw '{"limit": 3}' -o json 2>&1 | head -80)` then `linear list-comments --issue-id "abc1ceae-aaaa-bbbb-9aaa-6bef0325ebd0" 2>&1)`
So even the --help has filtering by default. Current models are pretty good
[0] https://github.com/steipete/mcporter
CLIHub
- written in go - zero-dependency binaries - cross-compilation built-in (works on all platforms) - supports OAuth2 w/ PKCE, S2S, Google SA, API key, basic, bearer. Can be extended further
MCPorter
- TS - huge dependency list - runtime dependency on bun - Auth supports OAuth + basic token - Has many features like SDK, daemons (for certain MCPs), auto config discovery etc.
MCPorter is more complete tbh. Has many nice to have features for advanced use cases.
My use case is simple. Does it generate a CLI that works? Mainly oauth is the blocker since that logic needs to be custom implemented to the CLI.
But MCP today isn’t ideal. I think we need to have some catalogs where the agents can fetch more information about MCP services instead of filling the context with not relevant noise.
The point is push vs pull.
I know I saw something about the Next.js devs experimenting with just dumping an entire index of doc files into AGENTS.md and it being used significantly more by Claude than any skills/tool call stuff.
I didn't release the website yet. I'll remove the link