pmkelly4444

8 ms·

1.
We gave an AI agent eyes. It didn't even use them(agentvoyagerproject.com)
2 points by pmkelly4444 4mo ago|0 comments
2.
Skills vs. MCP vs. prompts: which agent setup works best?(agentvoyagerproject.com)
3 points by pmkelly4444 4mo ago|1 comments
3.
The Slop KPI Era: How Tokenmaxxing Is Making AI Worse(portofcontext.com)
4 points by pmkelly4444 6mo ago|0 comments
4.
Show HN: Python Local Sandbox Code Execution (Podman and Uv)(github.com)
4 points by pmkelly4444 9mo ago|0 comments
5.
Pctx-Py – Code Mode for Python Tools and MCP(portofcontext.com)
2 points by pmkelly4444 10mo ago|0 comments
6.
A Python SDK That Saves AI Tokens with Code Mode(pypi.org)
1 points by pmkelly4444 10mo ago|1 comments
7.
by pmkelly4444 10mo ago
If you are building AI Agents in Python and want to improve token efficiency, please have a look at the RC of the pctx Python SDK
8.
by pmkelly4444 11mo ago
e2b.dev is focused on this space
9.
by pmkelly4444 11mo ago
yes! coming soon
10.
by pmkelly4444 11mo ago
this makes sense, we should support a model where the code snippets can all be stored on the filesystem rather than in the context window from the MCP response
11.
by pmkelly4444 11mo ago
got it, yes so currently this is built just for one user - one set of credentials, but passing user credentials through is something we want to add. thinking a native connection to cloud auth managers is the best way to do this (clerk, auth
12.
by pmkelly4444 11mo ago
thank you, will get a more detailed benchmark out soon!
13.
by pmkelly4444 11mo ago
no dependency injection at the moment... this is something we are exploring. adding dependencies would require rebuilding the execution runtime, which is something we want to open up in the framework soon
14.
by pmkelly4444 11mo ago
hey! the way it works is that the llm is first given snippets in typescript that tell it how to use the various MCP tools. it then can code, in typescript, and execute all of the tool calls in the deno sandbox. so yes, it can only execute j
15.
by pmkelly4444 11mo ago
thank you! on your second point, check out how we lock down the sandbox with a custom deno runtime! https://github.com/portofcontext/pctx/tree/main/crates/pctx_... on third, will def get some benchm
16.
by pmkelly4444 11mo ago
nice! will follow your progress! love that this runs locally as well
17.
Show HN: Pctx – OSS to build efficient AI agents with Code Mode(github.com)
1 points by pmkelly4444 11mo ago|0 comments
18.
Open Source and Local Code Mode MCP in Deno Sandboxes(portofcontext.com)
76 points by pmkelly4444 11mo ago|26 comments
19.
Code execution with MCP: Building more efficient agents(anthropic.com)
39 points by pmkelly4444 11mo ago|5 comments
20.
by pmkelly4444 1y ago
Interesting, are you using LLM rules files?
21.
by pmkelly4444 1y ago
And you're stuck maintaining the glue by hand because no generator understands the relationships between all these different systems.
22.
Show HN: Sideko – Hybrid deterministic/LLM generator for API SDKs and docs(github.com)
26 points by pmkelly4444 1y ago|5 comments
23.
Building an SDK Generator: Maintaining Custom Files(sideko.dev)
1 points by pmkelly4444 1y ago|1 comments
24.
by pmkelly4444 1y ago
An SDK generation solution that utilizes a macro! in Rust to maintain custom files added between code generations with zero configuration.
25.
Can APIs be beautiful?(sideko.dev)
1 points by pmkelly4444 1y ago|0 comments
26.
Should we still call it DevRel?(sideko.dev)
1 points by pmkelly4444 2y ago|1 comments
27.
by pmkelly4444 2y ago
SJ Morris, HubSpot's Developer Community lead, discusses the unique challenges of developer relations (DevRel) at "developer-plus" companies, where unlike public relations, devrel must constantly justify its value despite ser
28.
by pmkelly4444 3y ago
extremely useful
29.
Full Stack Dev with FastAPI(sideko.dev)
1 points by pmkelly4444 3y ago|0 comments
30.
by pmkelly4444 3y ago
this is very cool! I just tried using it, unfortunately, my NextJS app dir project makes most requests from the server side, so it was only capturing "posts" made from the client. Is there a way to run it from the server?

More ›