Rules/config file sprawl (CLAUDE.md, AGENTS.md, .cursorrules)

It's best to start with a relatively clean GEMINI.md too. Many configs optimized for Claude doesn't work well because Argon didn't have those same failure modes.

thanhhaimai · Oct 28, 2026 · comment in “Gemini 4 Carbon”

The language it uses – that aweful, useless typical LLM languages! No AGENTS.md instruction, no per session instruction, no just two-prompts-prior instruction seems to work, hell sometimes instruction regarding that in same prompt doesn't work.

crossroadsguy · Oct 17, 2026 · comment in “DeepSeek-v4.1-Flash-1017”

It's easy to say "fix model errors in AGENTS.md / the system prompt." But finding a way to prove the effect of that change is the hard part. It's so frustrating when a model says "no, your guidance was fine, I just didn't follow it, I'll do better next time..." It just makes me want to yell "No!

justinweiss · Sep 28, 2026 · comment in “Coding AI without deterministic outcomes”

From what I gathered memories essentially work/load into context in the same way as CLAUDE.md, but the way it writes them is really annoying.

aytigra · Sep 28, 2026 · comment in “Prompting Claude Opus 5.5”

What's also annoying is that Claude Code now spams "agents-md: no CLAUDE.md found; AGENTS.md loaded" on every startup.

silverwind · Sep 24, 2026 · comment in “Claude Code reads AGENTS.md only when telemetry is on [fixed]”

Mind, the change does not affect me at all, because when both instructions are present Claude continues to read CLAUDE.md and Codex continues to read AGENTS.md. But if Claude started preferring AGENTS.md, that would be mildly annoying.

Wowfunhappy · Sep 23, 2026 · comment in “Claude Code reads AGENTS.md only when telemetry is on [fixed]”

I was quite annoyed with Claude earlier this year when I had to maintain duplicate files or rely on symlinks. Seemed like a waste of time simply because Anthropic didn’t want to take the “L” and accept that AGENTS.md was the new standard.

perganomoly · Sep 19, 2026 · comment in “Claude Code now reads AGENTS.md if there is no Claude.md”

The standard is put in place by how many harnesses use a specific pattern, and that has been AGENTS.md for a long while now. Long enough that Claude Code not supporting it is really just an act of forcing people into a proprietary workflow, because of course it's in their interest to make moving away as painful as possible given how little moat the LLM models themselves have.

askonomm · Sep 19, 2026 · comment in “Claude Code now reads AGENTS.md if there is no Claude.md”

It favors deterministic scripts for process but launches them via LLM and monitors for exceptions; also, no memory system which also disqualifies Claude from being at the center of this. Memory plugins didn't work well because instructions start accumulating all through the chain of instruction files (AGENTS.md, memory, etc) and that gets very undeterministic very quickly.

browningstreet · Sep 17, 2026 · comment in “Claude Cowork and chat are now one Claude”

While the idea of improving codebase search and documentation has some merit, any modern coding agent harness does not have the problem of starting each session blankly and giving generic advice. The actual issue that I have experienced is that it will waste tokens re-reading the codebase over and over again to understand your query, although you can mitigate that by proper use of AGENTS.md and documentation.

joefourier · Sep 16, 2026 · comment in “Show HN: Friday – Self-hosted persistent memory for AI coding agents (MCP)”

3. Saying AGENTS.md can enforce anything at all is a pipe dream, I am struggling with this all the time, in a long session the tokens from AGENTS.md are so far in the back of the context and so diluted that LLM's "attention" doesn't register them.

aytigra · Sep 14, 2026 · comment in “A computational constitution to stop LLM agents from bricking servers”

Every IMPORTANT RULE WHICH MUST BE FOLLOWED STRICTLY to not to do something from my AGENTS.md and skills are violated without any hesitation from LLM. Glorified autocomplete doesn't work this way.

bvrmn · Sep 11, 2026 · comment in “Measuring the sloppiness of code”

Show HN: Sensez – helping coding agents catch their own code smells Hi HN — I’ve been frustrated by how many code smells coding agents produce - no matter how many instructions I put in the AGENTS.md, even the most capable models still reach for dict[str, Any], add multiple boolean flags to functions or hide logic in a heavy nested function that I can't easily unit test.

popov95s · Sep 2, 2026 · story in “Show HN: Sensez – helping coding agents catch their own code smells”

This works alright if I include in the actual prompt, but annoyingly it often doesn't in CLAUDE.md or memory.

sampullman · Aug 24, 2026 · comment in “My agent.md to improve LLM-assisted code quality”

It has this tendency, but usually a sentence about this in the CLAUDE.md helps avoid it. It's annoying though that you have to add that sentence.

bonoboTP · Aug 22, 2026 · comment in “Claudette: Make Claude stop talking like a BuzzFeed article”

Using AI agents is amazing, but in long-term, I started to have more and more problems about the context of them. They keep forgetting what we are working on, opening a new chat session clears all their context, my CLAUDE.md had like 1000 lines.

dat999zx · Aug 22, 2026 · story in “Show HN: Knowl – CLAUDE.md hit 1000 lines, so I built memory that prunes itself”

I've explicitly included instructions in my home CLAUDE.md to avoid this, because it's one of the most annoying things about how Claude writes comments.

vilos1611 · Aug 22, 2026 · comment in “Claudette: Make Claude stop talking like a BuzzFeed article”

Unfortunately, this doesn't work with CC's other features like the lazy loading of CLAUDE.md files in subdirectories.

postalcoder · Aug 19, 2026 · comment in “Feature Request: Support AGENTS.md”

And God, Claude is stupid: it keeps printing credentials despite its own guardrails, my prompts, AGENTS.md, etc. And what's worse, the OpenClaw uses MacOS Keychain which keeps asking for permission, making it extremely annoying and unuseful like the old Windows XP's "need administrator approval"

xinbenlv · Aug 19, 2026 · comment in “[dead]”

Everytime i resort to cursing it , it understand. Even putting `ASD-STE100` or simplified english in the claude.md doesn't work.

dustypotato · Aug 13, 2026 · comment in “DeepSeek V4 Pro 0813”

No amount of CLAUDE.md and rules tuning changed that. Is there a way?

physix · Aug 5, 2026 · story in “Ask HN: How to turn Claude Code into an effective software engineer”

it is so frustrating to me that claude code still does not understand how the bash tool properly functions and no matter how much I mention how it works in my CLAUDE.md it still misuses it.

zeratax · Aug 5, 2026 · comment in “Web Security is Too Hard”

Managing context in claude.md/agents.md files is fine but so much of the knowledge required doesn’t fit cleanly in one repo or the other. Plus, much of the team (product, design, support) does not work in repos but still needs a place to share context.

mat10112 · Jul 30, 2026 · story in “Show HN: Memsprout – share AI context with your teammates”

I was using Opus 4.6 until 2 days ago with no CLAUDE.md or anything and it was great. Tried out Opus 5 and it's been a super annoying experience out of the box.

mupuff1234 · Jul 29, 2026 · comment in “Interview with Boris Cherny [video]”

There are dozens more. The innate non-determinism of the LLMs flips the dice sometimes and that leads to subtle bugs -- which is maddening, especially if the disciplines on how to write one thing or another are clearly spelled out in `AGENTS.md`.

pdimitar · Jul 26, 2026 · comment in “The new rules of context engineering for Claude 5 generation models”

I ended up adding this to every one of my `claude.md` files across all my projects, really annoying tbh

rekpero · Jul 24, 2026 · comment in “Em dashes are amazing”

The most frustrating thing for me beyond the feedback black hole is that when things do inevitably go wrong, trying to point out the concrete issue and instructing the LLM to do better around it is challenging; a lot of the time directives (even those in a global CLAUDE.md!) are just ignored either outright or as the context grows, or fail to be passed down to subagents; negative prompts are discouraged because of t…

_verandaguy · Jul 24, 2026 · comment in “How do we stop vibe coding?”

Managing context in claude.md/agents.md files is fine but so much of the knowledge required doesn’t fit cleanly in one repo or the other. Plus, much of the team (product, design, support) does not work in repos but still needs a place to share context.

mat10112 · Jul 21, 2026 · story in “Show HN: Memsprout – share your AI context with teammates”

From my tests agents.md does NOT work with copilot.

phplovesong · Jun 8, 2026 · comment in “Do agents.md files help coding agents?”

I worry about the next generation. For me the mental framing of "It's all hallucinations, some of those hallucinations are useful" is helpful to keep frustration in check as I ask it to review the same implementation plan for the 4th time and it turns up different issues because the input was slightly different, or review the output code and see allow(dead_code) despite my claude.md forbidding it.

duttish · Jun 6, 2026 · comment in “Ask HN: Why is the HN crowd so anti-AI?”

With the additional of adding ClAUDE.md files describing correct patterns. However, besides the "regular", look at shit and clean it, is there any tools, or smarter ways to make our codebase easier for other to contribute to given the new tools avaialable?

kasnaka · May 20, 2026 · story in “Ask HN: How to make a mono-repo AI-Ready?”

What sold you on it? Mainly interested in: how well the agent actually uses the structured data, how painful setup/indexing is, and whether it's worth it over simpler approaches (ripgrep, good project docs, AGENTS.md).

fariswyatt · May 19, 2026 · story in “Ask HN: Is there a good code intelligence MCP server yet?”

Now, I'm using Claude or Codex (GPT-5.5) for frontend and backend and it just gets it right first time more often than not. I've been making use of things like LSPs, Context7 and CLAUDE.md (global and per-repo) and it just stops doing the dumb LLM things that I hate.

AussieWog93 · May 19, 2026 · comment in “The last six months in LLMs in five minutes”

There are really dumb ones where the bot writes 10 paragraphs about how he implemented the feature, but the entire changeset is adding one line to .gitignore or adding a CLAUDE.md file. There are even worse ones where the bot submits 3000 lines of code that seemingly works, but you have to spend an hour to figure out why it doesn't work.

xdennis · May 18, 2026 · comment in “We stopped AI bot spam in our GitHub repo using Git's –author flag”

I'm getting really tired of having new bugs pop up on me or subtle behavior change near daily that requires me to change things. The most annoying thing ever that was just introduced is a giant spew of context mode crap that Claude aggressively adds to every CLAUDE.md file, and I can't find a way to turn it off.

freedomben · May 16, 2026 · comment in “OpenClaw Creator Spent $1.3M on OpenAI Tokens in 30 Days”

The trigger was "look at this record" — I should have read that as "run a SQL query" and reached for tidewave immediately. If Opus 4.7 doesn't follow simple CLAUDE.md instructions, I'm not sure what benefits other markdown files could bring.

reedlaw · May 7, 2026 · comment in “Show HN: Agent-skills-eval – Test whether Agent Skills improve outputs”

README.md should be for humans CLAUDE.md for machines. Most agents use README as a storage for EVERYTHING related to the project by default, which is annoying for humans who just want to figure out a) what the thing does b) how to install it.

theshrike79 · May 2, 2026 · comment in “Apple accidentally left Claude.md files Apple Support app”

I’ve always struggled with what should be in Claude.md that doesn’t belong in readme.md or a similar supporting file.

data-ottawa · May 1, 2026 · comment in “Apple accidentally left Claude.md files Apple Support app”

It's incredibly frustrating when I've spelled out in CLAUDE.md that it should SSH to my dev server to investigate things I ask it to and it regularly stops working with a message of something like:

joshstrange · Apr 23, 2026 · comment in “An update on recent Claude Code quality reports”

I had to add extra instructions in Agents.md just to make it less snarky. I was annoyed enough that I explicitly told it not to use the word “canonical.”

chandureddyvari · Apr 12, 2026 · comment in “Pro Max 5x quota exhausted in 1.5 hours despite moderate usage”

The last two weeks with Claude has been a nightmare with code quality, it outright ignores standards (in CLAUDE.md).

zamalek · Apr 9, 2026 · comment in “Clean code in the age of coding agents”

Very obvious with the GEMINI.md. I wish there was a rule to ban AI submissions.

_lvbh · Apr 4, 2026 · comment in “Show HN: Mtproto.zig – High-performance Telegram proxy with DPI evasion”

If it does, the plan was insufficient and/or the context around the request (agents.md, lessons.md, or whatever tools and documents you use ) weren't sufficient. EDIT: Maybe it doesn't work in cursor, but I continue to use vscode to review diffs and dig in on changes.

criley2 · Apr 2, 2026 · comment in “Cursor 3”

4) claude decides it can't fix the problem and does a reset hard. I have this right at the top of my CLAUDE.md and it makes things better, but unlike codex, claude doesn't follow it to the letter.

jeswin · Mar 30, 2026 · comment in “Claude Code runs Git reset –hard origin/main against project repo every 10 mins”

The linter loop is the only thing that can keep the agent sane in my opinion, and if anyone can generalize bun+tsc loop to other tasks, this would finally be a way to trust LLMs output. I was annoyed at how Claude Code ignores my CLAUDE.md and skills, so I was looking for ways to expand type checking to them.

mifydev · Mar 26, 2026 · comment in “HyperAgents: Self-referential self-improving agents”

Also the process sounds like a nightmare: "it broke and I asked 4 different LLMs to fix it; my `AGENTS.md` file contained hundreds of special cases; etc." I thought this article was intended to be a horror story, not an advertisement

capitalsigma · Mar 15, 2026 · comment in “100 hour gap between a vibecoded prototype and a working product”

So, maintaining a fork is really not that serious of and endeavor anymore. I'm actually maintaining a fork of Zed with several additional features (Claude Code style skills and slash commands, as well as a global agents.md file, instead of the annoying rules library system, which I removed, as well as the ability to choose models for sub-agents instead of always inheriting the model from the parent thread; and yes, …

logicprog · Mar 10, 2026 · comment in “Redox OS has adopted a Certificate of Origin policy and a strict no-LLM policy”

Add one line to your CLAUDE.md pointing at that directory, and your agent starts every session already knowing your stack. Outcome to expect: fewer tool calls, fewer hallucinations about your specific setup, and lesser context to share every time.

TheBengaluruGuy · Mar 6, 2026 · story in “Show HN: Droidctx – generate Markdown docs from your prod infra for coding agent”

Show HN: Codebase-md – Creates Claude.md, .cursorrules, AGENTS.md from any repo Hey HN, I built codebase-md because I was tired of manually writing and maintaining

anandsaurav668 · Mar 6, 2026 · story in “Show HN: Codebase-md – Creates Claude.md, .cursorrules, AGENTS.md from any repo”

I was able to fine tune the agent client-side and then it worked until the context cleared, but I didn’t want to hard code into context/agents.md how to access an API that will keep changing. I hate all this non-deterministic programming stuff but it's still too good to not do it :)

dervishcat · Mar 5, 2026 · story in “Show HN: Stop LLMs from brute forcing (guessing) APIs”

It gets to the very heart of how CLAUDE.md works, and I apologize if my previous answer was ambiguous. Painful.

zahlman · Mar 4, 2026 · comment in “What Is Code Review For?”

I built this after getting tired of writing CLAUDE.md files by hand every time I started a new project or onboarded to an existing codebase.

vexorkai · Mar 1, 2026 · comment in “Show HN: Rulegen – Auto-generate Claude.md and .cursorrules from your codebase”

I did /init and now CLAUDE.md is on several layers. I wish there was a reverse init and minimum as needed init.

hbarka · Mar 1, 2026 · comment in “Switch to Claude without starting over”

You re-explain your architecture, conventions, past decisions. CLAUDE.md helps but it's manual and doesn't scale.

tstockham · Mar 1, 2026 · story in “Show HN: Engram – Memory for AI coding agents (2.5K installs, 80% on LOCOMO)”

Show HN: I got tired of syncing Claude/Gemini/AGENTS.md and .cursorrules

wolfejam · Feb 27, 2026 · story in “Show HN: I got tired of syncing Claude/Gemini/AGENTS.md and .cursorrules”

You need to kick off the sessions to read the ~/.claude/CLAUDE.md file. Annoying.

LouisvilleGeek · Feb 19, 2026 · comment in “Claude.md isn't always read and followed”

I spent the last few months obsessed with a specific bottleneck: every time I start a new AI agent project, I waste the first few weeks on the same "foundational" plumbing. Whether it is building lead research pipelines, setting up data extraction, or wrestling with .cursorrules to stop hallucinations, we are all rebuilding the same generic 80% over and over.

teemuki · Feb 10, 2026 · comment in “Skene Cookbook – 700 open source production-ready AI skills”

anything cloud-hosted means my codebase context leaves my machine. The setup cost was always too high and privacy never guaranteed, so I stuck with CLAUDE.md files.

rendro · Feb 8, 2026 · story in “Show HN: Sediment – Local semantic memory for AI agents (Rust, single binary)”

"I don’t get why people get worked up over MCP" Because we tried them and got burned? "to help us get more context into agents in a more standard way than everyone writing a million different markdown files and helper scripts." Agreed it's slightly annoying to add 'make sure to use this CLI/API for this purpose' in AGENTS.md but really not much.

bravura · Feb 7, 2026 · comment in “Introducing the Developer Knowledge API and MCP Server”

biggest annoyance is the over-defensiveness mentioned, and that Claude keeps trying to use Jason instead of JSON. Also, Claude has some bad habits around aliases that it does even though it's pretty explicitly mentioned in CLAUDE.md, other annoying things like doing `case functioncall() do nil -> ... end` instead of `if var = functioncall() do else`

dnautics · Jan 25, 2026 · comment in “150k lines of vibe coded Elixir: The good, the bad and the ugly”