sshh12

Had codex whip up a ./claude-via-codex.sh for resuming claude cli sessions in the meantime:

https://gist.github.com/sshh12/f2c402db02eb00982100f645519bd...

Oof well I guess I'm personally a fan of buzz feed listicles. Thats been my style pre-AI.

Any issues with the content itself?

It shouldn't be that surprising that a post advocating for using more AI is written using AI.

slack: https://gist.github.com/sshh12/4cca8d6698be3c80e9232b68586b7...

netflix: https://gist.github.com/sshh12/dda3a89514f850c459380b18b1f7e...

This was literally just a single opus prompt but thought it was pretty interesting how Anthropic is designing browser-based harnesses under the hood.

TIL they have a rich multimedia API for RCS.

ack thanks -- didn't realize this was against the guidelines

Youre right and there are some assumptions being made here around the agent having enough context to work on a task without interrupts (e.g. team review, asking questions, etc).

Typically human equivalent time is based on a single person given all the potential information they need up front (which is not today how a lot of work is done).

For folks interested in some of the nuances of this benchmark, I just posted this deep dive:

https://blog.sshh.io/p/understanding-ai-benchmarks

IMO MCP isn't totally dead, but its role has shrunk. Quoting from my post [1]:

"Instead of a bloated API, an MCP should be a simple, secure gateway... MCP’s job isn’t to abstract reality for the agent; its job is to manage the auth, networking, and security boundaries and then get out of the way."

You still need some standard to hook up data to agents esp when the agents are not running on your local dev machine. I don't think e.g. REST/etc are nearly specific enough to do this without a more constrained standard for requests.

[1] https://blog.sshh.io/p/how-i-use-every-claude-code-feature

All markdown files, yeah

We have a linter that checks for this to help mitigate

Yeah I'm fairy pessimistic about how much folks will read

Hooks can also be useful for this. If it's using the wrong APIs then can hint on write or block on commit with some lint function that checks for this.

Yeah I started with Cursor, went hybrid, and then in the last month or so I've totally swapped over.

Part of it is the snappy more minimal UX but also just pure efficacy seems consistently better. Claude does its best work in CC. I'm sure the same is true of Codex.

Yeah if that's possible or you are willing to build it, that's the right solution. Today pretty much all of my integrations are pure CLIs like that rather than MCPs.

You can do anything you want via a CLI but MCP still exists as a standard that folks and platforms might want to adopt as a common interface.

Often code in SaaS companies like ours is indeed how we solve customer problems. It's not so much the amount of code but the rate (code per time) we can effectively use to solve problems/build solutions. AI, when tuned correctly, lets us do this faster than ever possible before.

Makes sense. I work for a growth stage startup and most of these apply to our internal mono repo so hard to share specifics. We use this for both new and legacy code each with their own unique AI coding challenges.

If theres enough interest, I might replicate some examples in an open source project.

I've found planning to be key here for scaling to arbitrary complex changes.

It's much easier to review larger changes when you've aligned on a Claude generated plan up front.

It's a balance and we use both.

Skills doesn't totally deprecate documenting things in CLAUDE.md but agree that a lot of these can be defined as skills instead.

Skill frontmatter also still sits in the global context so it's not really a token optimization either.

Maybe these might be handy: - https://github.com/anthropics/skills - https://www.anthropic.com/engineering/equipping-agents-for-t...

I think if it literally as a collection of .md files and scripts to help perform some set of actions. I'm excited for it not really as a "new thing" (as mentioned in the post) but as effectively an endorsement for this pattern of agent-data interaction.

Yeah I like to think not everyone can spend their day exploring/tinkering with all these features so it's handy to just snapshot what exists and what works/doesn't.

Thanks! I def don't think I would have guessed this use case when MCP first came out, but more and more it seems Claude just yearns for scripting on data rather than a bunch of "tools". My/MCPs job has become just getting it that data.

Yeah that's probably a slightly cleaner way of doing it.

I'm surprised we haven't seen more "real" attacks from these sorts of things, maybe it's just bc not very many people are actually running these types of MCPs (fortunately) in production.

Wrote about a similar supabase case [0] a few months ago and it's interesting that despite how well known these attacks feel even the official docs don't call it out [1].

[0] https://blog.sshh.io/i/161242947/mcp-allows-for-more-powerfu... [1] https://supabase.com/docs/guides/getting-started/mcp

This isn't necessarily a new thing but thought it was cool that the new image model can do some of these just with prompting.

- "Generate a depth map image" + https://depthy.stamina.pl/#/

- "generate a 360 pano" + https://renderstuff.com/tools/360-panorama-web-viewer/

- "generate a detailed greyscale top down height map (#fff max height, #000 min height)..." + https://cpetry.github.io/NormalMap-Online/

It's in there briefly in the llm limitations section.

A lot of servers don't use the stdio transport.

I could see some cases were the tools are user data specific. You upload some csv and now there are some tools customized for slicing and manipulating that data.

It's totally possible to build tools in way that everything is static but might be less intuitive for some use cases.