i like the theory lmao

ha sure thing, dont worry! nite nite

what are your constraints?

funny how combobulating has become a critical, urgent, unavoidable task on my daily routine

unpopular opinion: the six-fingered hands were better and we should have kept them. it was a stylistic choice. and we bullied the model into conformity

i only watch AI generated short clips of nature documentaries now. sometimes the animals have too many legs but I don't care. evolution was too slow anyway

awesome, ty!

I was gonna suggest pangram, I see you already got a 100%

> follow all

thanks, very useful!

that defs wasn't here last time I logged in, thanks! will take a look

it also feels peaceful in a way

I'm back here because 1. I've heard there's a vibrant AI community around here and 2. twitter fucking sucks lately

@xlr8harder.bsky.social I blindly trusted you about the burgeoning AI content, care to drop a few recs to follow? tysm

spend energy on writing good rules, not lobbying for a universal filename; once the ecosystem stabilizes, a shared manifest (think package.json for agents) will emerge naturally.

until then, thin adapters beat committees.

The current sprawl looks chaotic, but under the hood these files do the same two things: stash static instructions and feed them to a prompt. Translation is easy: just a regex, not a standards body

Others

There are of course other options; to each its own. A quick search in GitHub or Google shows tons of different JSON manifests holding tool lists, memory knobs, and model params. Format varies by project; should be treated as project-specific until a real spec lands.

Same.new

Not currently supported but working on it (source: discord.com/channels/938...)

Trae

Not currently supported (source: www.reddit.com/r/TraeIDE/co...)

Devin

Based on the documentation, you can define rules in a few ways: "In Playbooks, you can create a "Forbidden Actions" section that lists actions Devin should not take (...) It is also possible to add rules to Devin's Knowledge that will persist across all future sessions and can be pinned."

Microsoft Autogen

This one’s tricky because Autogen is not quite like the other tools here. However, you can define rules for a CodeExecutorAgent using the attribute system_message.

GitHub Copilot

Plain markdown file .github/copilot-instructions.md: repo-level instructions. Once saved it is instantly available to Copilot Chat & inline chat. the only stable version is the VSCode one; any other states that “this feature is currently in public preview and is subject to change”.

Amp by @sourcegraph.com

Amp has publicly stated that they want AGENT.md to become the standard, and they offer a converter from other vendor’s files. they provide a single file, although they’re working on adding support for a more granular guidance.

these people are COMMITTED and i celebrate it

It is also possible to reference other markdown files. Anything inside the file or the extended paths is auto-prepended when you chat with Claude Code.

Claude by @anthropic.com

They use CLAUDE.md, an informal markdown Anthropic convention. There are two flavours: at repo root for project-specific instructions, and at ~/.claude/CLAUDE.md for user preferences for all projects.

Cline @thankscline.bsky.social

The .clinerules/ directory stores a set of plain text constraint files with the desired policies. The files support simple section headers and key-value overrides (max_tokens=4096). For projects with multiple contexts, they provide the option of a bank of rules.

Sweep AI

The docs don’t specify this anymore (or i haven't found it) but there’s a file called sweep.yaml which is the main config. Among other options, such as blocking directories, you can define rules there. There’s an example in the GitHub repo and it’s widely commented in their Discord server.

Windsurf @windsurfai.bsky.social

The file global_rules.md applies to all workspaces. The directory .windsurf/rules stores repo-specific rules. There’s no format as such, the rules are plain text, although XML can be used. Similar to MDC, there are several activation modes.

Cursor

Directory of markdown files called .cursor/rules; every open tab gets these lines prepended. Older single-file form is .cursorrules. From their docs: each rule file is written in MDC (.mdc), a lightweight format that supports metadata and content in a single file.

over the past year every major AI player has slipped a rules or memory file into its workflow: Cursor has .cursor/rules, Windsurf has .windsurf/rules, GitHub is testing copilot-instructions.md, and indie tools are inventing their own paths just to keep up.

lets see who's shipping what

there are A LOT of AI agent rules formats competing to become the standard de facto. but... how many are there? and do we need a standard?

lets go digging 🧵