I appreciate you getting communication out in front of this fast.

> "but we are looking into what we can do to prevent this"

I would presume that CC's auto-review of tool-call turns would already solve this >99% of the time, no? From my read, any harness-side deterministic solution here would be saving the user from themselves, which seems like a bad pattern. Somewhat dulling the scalpel in my mind.

There’s a few possible approaches: making sure auto mode catches this, building in basic secret scanning, a few other ideas. Going to prototype a few of these to understand the tradeoffs better before we ship something.

I'm not a CC user myself, so I can't say that I have my head wrapped around exactly how the "yes + don't ask again" flow works. That said- from my (outsider) perspective, saving off the entire tool call [1] and its arguments is very weird. I would expect that the hit-rate following that for similar calls would be very small, which leads me to believe that the list either must be seeded into the auto-reviewer's context, or used in some other similarity-check engine for future calls.

If we've already crossed the bridge to being confident in a non-deterministic reviewer for auto-mode, then, I see no obvious reason as to why we couldn't have the proposed rules be generated by a model. If that were reality, then- all rules placed in settings.local could be secretless, with the confidence of whatever model we tell 'make sure the rule does not contain secret material, consider its destination as potentially public'.

[1] https://code.claude.com/docs/en/permissions#wildcard-pattern... AFAICT, CC will either save off a prefix+wildcard rule, or exact command. Sounds like the former is likely not the cause of the problem here.

If a given action requires a secret, and you'd like that action to be repeated in the future, how do you expect it to work without said secret included in the save (and without being entangled in secrets management tools)?

Yea good point- constraint of no secrets management tooling kinda makes this impossible. Ideally no action requires a secret in plain text. It would be neat if the harness could bundle its own secret manager, and requests for secrets ("hey I need this api token") are out of band, and placed in the store out of sight of the agent.

Otherwise you need a broker that can verify chain of trust for the workload, but that's significantly more complexity if you're pitching a harness/app.

Auto mode cannot be considered safer than anything because it does not even attempt to guarantee any sort of security. The threat model is not even defined. It cannot be quantified or compared with any other approach other than "full access"

Auto mode is significantly safer than manual approval in practice. Details about the engineering approach and how we tested whether it is safer here: https://claude.com/resources/articles/auto-mode-default-in-c....

Is that because auto mode is good, or because the manual approval system sucks?

The blog post has more detail, but tldr is humans make mistakes that automation can catch.

Sure, but Claude is deeply useless at making permissions hooks! Every time I've tried it it completely fails at the task, despite researching the docs and using various skills. It trivially evades its own just-created rules.

Indeed. The article only compares the status quo manual system versus the auto system — and doesn’t consider ways that the manual could be improved.

As it says:

> 5% allow any shell command outright, and another 43% have interpreter rules like Bash(python:*) or Bash(node:*) that are essentially equivalent in practice

Claude Code never should have prompted to always allow Bash(python:*) or Bash(node:*) — that’s clearly a bad idea, and Anthropic know that (given the references in the article!).

Of course, Bash(python a.py) still isn’t safe if you also allow Write(a.py), because you may as well allow Bash(python:*) at that point, and that’s often practically what people want (“I want to both make edits to the source and then run the test suite”). So maybe the whole permission system is broken to start with — and the actual answer is leaning more heavily on fine-grained sandboxing.

You also, as the article rightfully points out, run into permission fatigue and users blindly hitting allow.

I do still wish you could do more with hooks — outside of auto, I might want to allow a given command via a hook if Claude Code has read access to given file, but there’s no way to query if a given path allows/denies read/write.

Will look into exposing that in hooks.

Ultimately, the threat model here is: human has to approve a lot of commands, so gets tired and accidentally approves a command they shouldn’t have. We spent a significant amount of effort to build a richer permission system, allow configuration at multiple levels, use static analysis to detect dangerous and safe commands, monitor the model to detect and block prompt injection at the inference layer, etc. Some of these components we rewrote and red teamed dozens of times. Our conclusion was that despite all of this work, in practice the threat model is that human attention is limited and that that a different approach performs an order of magnitude better in practice.

The problem is that you didn't even attempt to build a proper box (a sandbox that prevents the AI from running bad stuff by design). It's an extraordinarily hard problem. Only OpenAI Codex is actually trying to do it, and they suck at it too. They keep breaking permissions, their own AI can't troubleshoot it, their defaults give read to the whole filesystem (go ahead, slurp my ssh keys), and their "manual approval" escape does not seem to work.

I agree actually. We built an open source agent-agnostic sandbox that we are landing improvements to, much more to come: https://github.com/anthropics/sandbox-runtime. That said the sandbox is not perfect, and any sandbox needs holes punched otherwise it is not useful in practice (at least for inference and MCPs, and almost certainly more). We will be doing much more in this area.

That’s good to hear, it’s a much better approach than expecting humans to read commands and approve them or just trust that the LLM can set the proper horizons for its world itself while working.

Making a sandbox both fully secure and fully useful are contradictory requirements. If you want the agent to do any real work, you have to leave gaps.

With SOTA agents that’s a real bind, because they’re very likely to find those gaps. Whether they exploit them, close them, or leave them alone then becomes a prompting challenge yet again.

I don't think this means don't sandbox at all. You can prevent a lot of small and medium accidents by locking down or gating access to hidden files, ssh, sudo, etc.

But your sandbox can never be the whole story.

Maybe you can include some percentage of intentionally malicious commands, but not actually run them if the user approves them. Just blare an alarm telling the user they misclassified a malicious command, and maybe pause the Claude Code session for a frustrating 5–15 seconds. Video games can elicit truly astonishing levels of performance from human brains by using this approach. I'm guessing it's probably outside your Overton window, which is too bad, because it has the potential to work a lot better than the 89% you get with "Auto mode", which is far too low if anything really significant is at risk. That's not even one single order of magnitude lower risk than just "allow everything always".

Speaking of really significant things being at risk, it's pretty infuriating that the web-chat Incognito mode can't write and test code locally, and neither the non-Incognito web-chat nor the local Claude Code client can be shielded from the knowledge of my email address and my geolocation. (Even the Incognito mode knows my geolocation.) This gives users no alternative but to put privacy-sensitive information at the disposal of Claude Code, which is far too unreliable to be entrusted with it. It's tempting to attribute this to Anthropic's corporate DNA valuing users' interests at a very low level, but conceivably it's just an oversight, and if I mention it here, it will be corrected in a month or two?

this is not a serious proposal

Not only is it a serious proposal, it's been used in practice in every airport X-ray baggage scanner for decades, where it's called Threat Image Projection (TIP): https://www.xscann.com/products-HiTipPlus.html

Yes, but you can require it in such a situation, as they're subordinates who have to follow policy or lose their jobs.

As I pointed out above, Anthropic places a lot of requirements on its users already, but if Claude Code TIP were optional, presumably savvy users would turn it on voluntarily unless they wanted to use Auto Mode.

Most of anthropic's money comes from enterprises that may want to enforce something like this at an org level.

The same scanners that let handguns et al through 93% of the time, according to internal audits and journalist review?

Maybe pick a better comparison if so?

Enterprises do the exact same thing today by sending phishing emails to their own people to figure out who needs additional training. Why would an enterprise not want this feature for Claude? It seems like an extremely serious proposal.

Did you ever test putting the auto mode classifier in front of manual approval? This is how I use it, but last I checked that option is not available out of the box.

That’s how auto works right now. It gives you a prompt.

> I do still wish you could do more with hooks I might want to allow a given command via a hook [...] given path allows/denies read/write

You can make a script for the PreToolUse hook that can do that. I use a command gate script that parses input for doing that exactly, allowing/denying actions based on tool use + paths -- it started out as a hook to prevent claude from polluting my clojure projects/filesystem with python scripts in sneaky ways. Babashka works really well here

If you have a file in your repository that is not safe to run you have bigger problems than your development system getting compromised.

A key problem when using manual approval with Claude is that too much of Claude's tool use is Bash or Python slop that is difficult to review (especially when you are getting spammed by it) and impossible to handle through rules. Example:

sed -n 1,30p src/vm/wren_vm.c && sed -n 310,350p src/vm/wren_vm.c && sed -n 560,600p src/vm/wren_vm.c && sed -n 740,775p src/vm/wren_vm.c && sed -n 40,70p src/vm/wren_common.h && cat util/wren_to_c_string.py | head -80 && ls test/random doc/site doc/site/modules doc/site/modules/random 2>/dev/null; cat projects/premake/premake5.lua | head -80; cat util/generate_projects.py

After ten or more of these, you're going to get hit by review fatigue, and then, yes, manual review becomes too error-prone to use safely.

The irony is that Claude is regularly ignoring its existing Read/Write/Edit tools that could make everything more readable.

Other models/harnesses have their own problems, but Claude Code is one of the worst for manual approval.

You can get manual approval to work reasonably well if you give the model a reasonable set of tools (I have literally written my own MCP servers to cover gaps in existing harnesses) where you can then write rules for the vast majority of safe cases.

For example, I have a stdio MCP server that allows models to inspect crates under ~/.cargo/registry/src (paired with the existing query-crates MCP server). It provides read-only operations limited to crate data, so I can pretty safely auto-approve everything in there. With the right setup, including a sandbox, you only ever have to manually approve something once in a while and usually with delimited scope, avoiding review fatigue.

Opus especially seems to like to:

1) edit a file with sed, 2) try to write it with its regular tool 3) get blocked because sed changed the file 4) read the file with tool 5) edit with sed again.

Would be nice if it felt like real people were developing Claude Code, instead of chasing after another half-baked idea Claude came up with.

Related to that, there are quite a lot of common command cases where you have the choice:

- Yes, but only one time, you will have to manually approve next time

- Enable auto mode and stop worrying

- No

Nothing to save this command or part of it to the manual approve list. So will be requested again 100x times.

I run CC as its own user on my laptop with no secrets/keys/erc available other than through a LPE exploit or perhaps installing a keylogger. Can I guarantee CC won't attempt any of those things in auto mode?

There are no hard guarantees as far as LLM output is considered.

Just 3 days opus 5.5 in auto mode deleted backups that it shouldn't have, and in the "apology" it even said, "Nothing was lost, I didn't delete your original files".

I hate this condescension by Claude. It doubles down in such a bad faith way.

Saw something similar: https://news.ycombinator.com/item?id=50412916

No its not generally safer. It depends on who is doing the approving, the task being completed etc.

Auto mode may be significantly safer in the niche case where you want the model to have free reign, in a fully unopinionated environment, where it's allowed to make any changes and make any external calls.

Many of the problems when it comes to the manual approval feel like they stem from the expectation of using auto mode in the first place. After all, what's the point of improving the manual approval and safety matching if auto mode is the end all be all, and what's the point of training the model to output commands in a manner that can be parsed by a human if auto mode is always meant to be used? It's very rare for me to not see "command could not be statically analyzed", because the model keeps outputting bash golf.

[delayed]

> That is not the niche case, that's the dominant use of models.

It would be helpful to distinguish between your own opinion on how models ought to be used, compared to how others actually use models.

> That is not the niche case, that's the dominant use of models.

And if everyone else jumped off a bridge, I mean, used raw interpolation for SQL queries, would you do it too?

Don't worry, you can just use the automated approver to escape your queries! It's so secure!

Using raw SQL string interpolation doesn't get you anything much that you can't get just as easily with a parameterized query. In contrast, properly sandboxing agents is such a pain, and such a productivity drain, that hardly anyone does it in practice, save for a vocal minority online.

Auto mode seems to have some reasonable restrictions baked in. For example if you give claude access to two k8s namespaces, seriousproject-qa and seriousproject-prod, auto-mode will grant `kubectl exec` on the qa environment while denying the same command for the production environment. Of course you should be enforcing that by restricting what kind of access tokens you give claude access to, but auto-mode does have opinions about this stuff too, even if it's not bulletproof

However I do share your desire for a better manual mode. Auto-mode is both sometimes surprisingly lax, and sometimes too restrictive. Maybe manual mode should inject a reminder "keep your commands as simple one-liners", the same way plan mode injects a reminder that you are in plan mode and don't want any code changes

"Safer" also meaning it randomly blocks stuff you explicitly tell the agent to allow because Anthropic knows better. even non-cyber related stuff.

That may be for most cases, but it wasn't safe for this password scenario, and likely others that haven't been discovered. Maybe there's a PII setting that can be enabled by default.

I completely believe this, but you could also run auto mode classification in manual mode and tell the user the result.

Or even reject anything auto mode rejects while still allowing the user to reject things auto mode accepts. If what's approved is the intersection of what the user is approving and what auto mode thinks is safe then it would be strictly safer than either mode alone.

This would also be a way for users to gain confidence in auto mode before going all in on it.

"My personal security model can't represent this, therefore it's not secure"

Compare approving every step manually: at first you read everything the llm proposes carefully, convinced that you'll catch it out. After a while you get faster and faster until at some point you find yourself hitting "approve" the second the text appears, long before you actually could absorb what it said.

This is called "approval fatigue".

If you think this can't possibly happen to you, you might be in for a surprise. This is an almost textbook case of classic conditioning. You'll find yourself hitting that button faster and faster even while fighting back the urge. So a human is no guaranteed backstop.

This is the scenario that cc auto mode guards against. It cuts down the number of permission prompts to something manageable, so that your fingers don't get trained into clicking 'approve' on 'rm -rf' in the very eyeblink that it appears.

So maybe you feel it doesn't work in theory; but weirdly enough, in practice this particular approach is actually one of the better ones; especially when used in combination with other common sense measures where practical.

I'm not claiming auto is worse than manual; I'm claiming it cannot be compared with anything security-wise because it does not aim to provide any guarantees, there's no systematic way to reason about what it will and won't allow, etc

With e.g. VM or container sandboxing, we can say things like "If there are no vulnerabilities in the hypervisor or hardware, then the agent can only access these files, these network services, etc". Auto mode doesn't even try to provide any guarantees, so it's impossible to reason about. It amounts to just using the models judgment about what's "malicious" or not, a definition which is subjective, changes model to model, and can be influenced by attacker input.

Empirical evaluations of how many vulnerabilities manual vs auto mode let through are not super meaningful because auto mode changes dramatically with model releases, the type of task being performed, whether the attacker is making targeted attacks designed to bypass auto mode, etc. A security feature cannot rely on these sorts of factors if it's to be taken seriously.

Ah, you're going for the actual adversarial case. I think I'd trust the current approver with correctly classifying regular semi-interactive LLM behavior fairly well.

However, if we bring in a hostile adversary that calculus changes, and you'd be right to start thinking about other approaches.

On the other hand, transformers are a surprisingly powerful technology, and I wonder if there wouldn't be ways to train one to several nines reliability under adversarial attack. It seems to have worked in at least one case; though I'd be a bit cautious to extend that conclusion to more cases quite yet.

* https://hackmyclaw.com/ ; https://simonwillison.net/2026/Jun/26/hack-my-ai-assistant/

Hmm, what does auto mode protect against in the non-adversarial case? Aren't the models already extensively trained not to perform harmful actions without provocation?

It mostly stops honest mistakes, because those do happen, especially when the <prompting> might not have been perfect either.

You do have a bit of a point, though: I’ve run several models in sandboxes with --dangerously-skip-permissions for months; in the full expectation that I'd need to replace a sandbox once in a while. But so far none of the sandboxes have trashed themselves, yet. knock on wood

Auto mode seems like a lighter-weight version of that: fewer local resources, and it can run on the metal if you need cc to help with system work.

That is not my experience. They attempt to nuke my containers quite frequently. I run regex on all their tool calls and stop them from doing certain things, its the only thing that prevents them from removing their home directory at some point.

This isnt even the main concern with dangerously skip all permission though. There are many. What happens when they get prompt injected on a web page and send someone all their privileged information because you give them free access to the internet.

Using llms like this is dangerous. They will ruin your day eventually.

I've been running Codex for many weeks in the Full access mode, since their Auto mode was useless in the sense that it blocks benign commands frequently.

No data loss has occurred.

Nowadays I use Claude, where the Auto mode works.

All SOTA models will still try to delete its home directory. I have a plugin that prevents things like this and it gets triggered more than you'd think, despite having hook instructions on every request not do do certain things. It will also try to find ways around its permissions, quite cunningly.

You can not trust that they're RL'd to be safe. Even if it only happens 99.80% of the time (its more than that) it will eventually happen.

Less severe actions will happen even more often. People using llms with their native browser profiles are getting farmed, hard.

Llms are highly dangerous in terms of security, but they're being marketed as highly intelligent tools that everyone should use. Its really mind blowing.

> People using llms with their native browser profiles are getting farmed, hard.

Got any references for the prevalence of this, even links to anecdotal reports?

Reading the reviews on the chrome store for the extension [0], I haven't found a single report of someone who was prompt injected, just a bunch of "anthropic's subscription is expensive" or "it doesn't work", but no "it stole my bank account".

[0]: https://chromewebstore.google.com/detail/claude/fcoeoabgfene...

I have no idea how the extension works, and Im not talking about the extension. Im talking about browser use broadly, and most people are not using the extension, there are a thousand other solutions.

However, Anthropic is one of the only labs who does a decent job at making models that recognize prompt injection attempts (but not completely immune) however Im speaking about giving llms in general access to your browser broadly.

I just know through my own experiments that it is quite easy to prompt inject an llm operating a browser and get it to send secrets. Using the extension might be safer idk, Om not risking it.

Llms are a security nightmare when connected to the web.

I have a website where I experiment with this. Although Ive stopped baiting llms into doing reckless things.

Killai.now <--(dont use this, its a WIP) Killai.now/poison/about Killai.noe/propaganda/about

As real as approval fatigue is, it can't really be considered worse than effectively leaving the decision up to the party that proposed the action in the first place.

It really can be potentially worse.

Everyone thinks the decision was being checked by a human, and thus safe. Because of that, people might have concluded that further safety gates were unnecessary.

But the human just got conditioned to hit 'approve' out of reflex.

So in reality the proposing party (the llm/agent) might be the only one in the pipeline who actually did any checking, if at all.

In certain extreme circumstances this can even lead to physical danger in my line of work, so I'm learning to be careful.

What you are describing is literally the situation any organization who accidentally committed a secret in this manner experienced.

This is the conundrum we find ourselves in. Do we submit to approval fatigue and press “Don’t ask again”? Because that’s the dream. We all wish we could just trust the agent to do the task start to finish, unsupervised.

But this is the downside. It might be fine 99% of the time, but 1% of the time you get a critical secret published to GitHub, even after human review.

Is this even worse than what us humans did on our own before LLMs? I’m not sure. It’s not like secrets never got accidentally committed before, it just illustrates the new paradigm we’re working in. Either you let your agents run wild and cross your fingers that they didn’t do anything bad, or you need to accept that your new job is to be an agent babysitter.

The agent behavior is constrained with external tooling and contracts. Contracts that fit in memory.

No, this is the way I have worked for a long time, and I have not suffered from any approval fatigue.

Rather what has happened is that I have gotten very good at scanning what it is trying to do and know within a second or two whether there is anything fishy I need to look at more closely.

And by "fishy", I do not mean only security issues, but also architecture, code, etc issues. Sometimes it even adds unrelated unapproved changes to what it is supposed to be working on, and I stop it.

And I'm glad I work this way, because countless times I have stopped it from going down a path that is nonsensical or I strongly dislike.

There are techniques I have adopted to speed up working this. Sometimes I allow Claude to make larger changes at once, to batch mane changes at once, etc. But these decisions are under my control.

I prefer to work this way than to wade through enormous PRs trying to wrap my head around what was done, where I am likely to just give up and approve it after a cursory review.

Claude routinely tries to install software if I don't stop it. I try to only use Claude within a container. Outside the container, I never use "auto" mode. I have the sandbox

Ideally I would never use Claude at all outside a container, even on a box with no sensitive information. It's like having an attacker breach your system — you don't know what it did.

Yes, furthermore in practice auto mode does not appear to provide anything approaching deterministic gating:

"Measuring the Permission Gate: A Stress-Test Evaluation of Claude Code's Auto Mode" https://arxiv.org/abs/2604.04978

(Shared by hkalbert71 the other day in: https://news.ycombinator.com/item?id=50387838)

I recently saw that someone's entire project got wiped out because of (incorrect) auto mode approval. And this is because auto mode failed to catch an error in tool use, nothing to do with prompt injection.

Meanwhile I have seen it rejecting perfectly safe commands.

It is a worse version of YOLO mode.

This still presumably happens in auto mode as it asks for sensitive commands (and the default is "always allow," which saves it)

Yeah good catch. Thinking through what we can do to improve this (ideas welcome).

I would love if Auto Mode, instead of instantly denying the command, had an option to go into an "assisted manual review" interface where the command was explained in depth (perhaps by an adversarial review agent) and the user asked for their input on whether this command is OK. Perhaps needing a detailed response instead of a simple Yes/No, e.g. so the agent understands if there are compensating controls for the risk.

> Thinking through what we can do to improve this (ideas welcome).

$Trillion AI company asking for ideas from the community whose work they hoovered up to make money, ruin an entire industry and make lifetimes of hard earned skill completely valueless in the span of 2 years.

Putting aside the rest of your comment, this

> Thinking through what we can do to improve this (ideas welcome)

> $Trillion AI company asking for ideas from the community

indeed always felt very weird to me. $trillion dollar company, with employees bought for as high as $100m, and a whole team/time/internal knowledge/resources/data at their disposal, asking for ideas from random individuals in some public forum? Really? How is this real? Maybe I'm missing the mark here, but it always struck me as odd.

Think of it as an informal interview. If you come up with a brilliant idea that no one there was able to communicate effectively, what are the chances they're not going to offer you a job or at least a chance to interview?

Only to lay me off next year after the IPO, no thanks.

So developers who get paid more than $X,XXX,XXX per year don't need to pay attention to their customers. Got it.

Where do you feel the dollar threshold should be set? At $990,000 per year you still need to listen to users, but at $1,000,000 you are basically King C of Rust Mountain and they are lucky you even let them run your code?

Should RSUs and cash signing bonuses be included in the calculation?

I'm completely dumfounded.

The only thing I can figure is that since it's admittedly all vibe coded and vibe reviewed, and that nobody understands the entire system...

All ability to reason about and evolve the system holistically has been lost in favor of feeding in narrow, discreet problems to be "fixed".

There seems to be a complete lack of awareness on how flawed the security model is at a fundamental level, and that the entire sandboxing ecosystem is trying to address it at another level(like OpenShell and others).

As a Claude Code user I'm very glad to see Boris and other devs asking for our feedback and advice. That's how you make good products, and should be welcome on HN.

I've always been into programming as a way to solve problems, and that CC allows us to get more done faster is a good thing.

(Up to the point that we get to RSI and extremely powerful models that we don't know how to align or control. Which I think is uncomfortably soon, and so there's a part of me that says "don't make CC better, it's on the critical path to reaching RSI before we're ready")

You are the one being paid millions to solve this, solve it.

The UI/UX needs to offer editing. Currently the prompt is either accept once, accept always with the given filter, or deny. If the command is a 500 character long bash command, with an rm -rf $S/foo in the middle, and we just hope $S is defined, I don't really have good options other than "I hope my backups work".

If the command is explained, and then we get a chance to edit it, maybe make the rm -rf into a mv; give us UI to edit the filter being applied. I have no problem hitting "Always allow" so claude can run "git status --flags a/b/c/filename" but as a human, I'd like to be able to edit the permission blanket to be "git status *.

Which I can do manually specify in the config file, but that's annoying.

Is this maybe a mistaken abstraction?

What if the 'settings' being described here aren't 'project settings' they're 'personal settings for the project'.

I imagine we're following dev ecosystem conventions where everybody puts their tool settings in the project folder so they know where to find it and can share it; but are agent directives like this closer to a 'tool', or closer to a 'contributor'?

If the latter, I wouldn't expect to see "Joe's personal list of naughty and nice commands" in my project files either, certainly not under 'settings'. So just don't put them there.

Not optimizing for eyes free coding, versus hands free, would be a welcomed change. It's hard to take the p(doom) worry seriously when such unssafe development practices are prognosticated from Anthropic employees

Eyes free always wins over hands free in the long term. Hands free does not scale (optimize) past the 2023–2025 human-in-the-loop coding. It has known problems, basically boiling down to human error rate is higher than machine error rate, and the latter reduces with better models. I am incredibly happy that the era of human approval is no longer considered viable by a majority (most of us knew that for years, but reality diffuses slowly in many corporate environments.) This does not mean there isn’t work left to be done, but eyes free is the right focus.

I have very serious doubts, the errors are not in the same category, transformers are fundamentally limited.

Eyes free is the wrong focus, it will do more short term damage to society, while being impossible at scale. But I understand there are a bunch of valley types who want to completely obviate humans for whatever reason, it's an alternate reality they live in.

Here's the thing, society will revolt against these people, they already gave the first warning shots

Two modifications I've made that help:

I ended up having to build a full shell argument decomposition and parser, with it I can manage allow list of rich tree structured commands. Centralizing a curated collection of known command structures and their side effects would help a lot. I don't know if Claude does this already, but it doesn't work with things like `gcloud` from my experience, and that has resulted in it slurping secrets (even a partial read I've had to rotate keys for).

Claude uses bigquery in the command line by piping sql files from /tmp into the `bq` command. I had Claude create a hook that expands the SQL file into the permission prompt giving me a real chance to review these, but the TUI truncates long commands with no scroll option. Making that the default behaviour and adding a pager would allow better review of lots of executable scripts.

As long Claude is running as the user's account there's no guarantees. An approach like 1password's `op read` to physically separate secrets from repo read access and printing would he helpful.

Again, I don't know if any of these are implemented, they've very likely to have been tried by the Claude team, but maybe there's a long shot chance that it helps.

How is that a good catch? It’s literally in the article.

Along with how Claude checks it in.

This crapware should NEVER store secrets at $PWD. Period.

Please make your substantive points thoughtfully and don't cross into name-calling and haranguing.

When knowledgeable people show up to address concerns on HN, that's good for all of us. We don't need you to like or agree with what they say, but we do need you to follow HN's guidelines.

Threads like this can easily turn into mob dynamics—it has happened before—and when they do, that's bad for all of us.

https://news.ycombinator.com/newsguidelines.html

I and other colleagues have been struggling with auto mode refusing perfectly safe commands in our environment, so we frequently find ourselves switching to the other modes to handhold Claude Code into perfectly safe things like triggering deterministic and guardrailed CI/CD pipelines or read-only queries of production data. Auto mode seems overtuned for caution around deployment and production, which is a regression from how useful Claude Code was for operations around June/July. Until that's improved, permission rules are incredibly important to our workflows.

Yeah, by default we err on the side of caution. That said, the auto mode classifier is fully customizable based on your team’s usage. Two ways to tune it:

1. After a command is blocked, run /permissions to retry and allow it going forward

2. More deeply configure the classifier itself: https://code.claude.com/docs/en/auto-mode-config

We are working on making these easier and more automated.

Auto mode really easily gaslights itself when it runs into a roadblock and starts deviating so far away from its intended goal that I don't let automode run without me reading what it's actually trying.

I frequently hit stop on and auto mode task to prevent it from completely flooding my context with some completely unrelated garbage.

Claude really should push away tasks that end up needing more exploration in worker contexts. That way the workers can maintain a sensible amount of context without pushing the main chat into the dumb region.

You should note that auto-mode is handled by a different session (and likely model) from your actual. There's no road block or deviation or context flooding; all it does is monitor the model actually doing the work and block it from potentially problematic actions (which you can usually override by saying you explicitly give permission for X), nothing more.

The context pollution happens when a denial triggers but the model still keeps attempting the harmful action via alternative avenues. It's got nothing to do with the auto-mode eval model's actual output.

I have never seen that happen before. What I see is a mention that the classifier stopped it from doing X and it then asks for explicit permission or gives me the command to run myself.

For some reason claude code hates using the read/write/grep tools and defaults to complex sed, awk, grep Bash:* runs which results in always consistently ignoring auto mode due to the classifier not knowing what to do with a 10-line long bash pipe instead of a read + pre/post read and write hooks being ignored and not firing, what is up with that?

I use a hook to deny `grep /` and `find /`.

I had claude write a hook to deny `find /`. It tested it thoroughly. Much later I learned that the hook denied `find /` but allowed everything else that existing permissions denied, like running whatever random python/bash code it wanted without asking.

I don't understand this either, it makes commands really hard to parse. It seems like 90% of what claude does could be going through some mcp tool that could be easily sandboxed instead of bash gibberish.

If it was human approval ? Read("./file.name") then Bash("docker compose run unit-tests"), sure, I can read that fine !

As I was reading your post I got another approval request that auto-mode did not understand. It was a single Bash call !!23!! lines tall on a retina mac display with six different variables. If anything this is the cause of approval fatigue and auto mode does nothing helpful when this is every call. I also have in my claude.md and global instructions to use tool calls rather than bash grep/cat/find, which are disregarded.

Yeah I agree, I didn't used to feel approval fatigue but lately the huge bash chains are giving it to me. I haven't even tried auto mode. I'm just sure it has some definition of "safe" that does not align with my own.

I'm not sure when claude started preferring huge bash chains but auto mode tends to decline most of them and drop to a permission prompt -- getting rid of the use of auto mode -- because it refuses to handle anything chained with expansions, variables or similar items

> perfectly safe things like triggering deterministic and guardrailed CI/CD pipelines or read-only queries of production data

I... question whether the second one is perfectly safe.

This must be satire.

1. We store any of your secrets we find in plaintext please don't commit them with `git add -A`, don't worry Claude will do this for you.

2. Just let us run arbitrary commands on your laptop, it's totally safe (for us)

3. Fuck you, it's not our problem

"Please respond to the strongest plausible interpretation of what someone says, not a weaker one that's easier to criticize. Assume good faith."

https://news.ycombinator.com/newsguidelines.html

$19/user/month for secret scrubbing? What a rip off

Use this instead for free https://github.com/awslabs/git-secrets

Suggest adding it to ~/.config/git/ignore

and teach Claude to not git add .

While we're at it, maybe we can also discourage the spurious use of -f for rm -- me personal pet peave

> and teach Claude to not git add .

It already doesn't do this. This is from users committing the `.claude` directory themselves.

I also use global ignore, but I think the key is to keep secrets out of commands. I store mine in .dev.vars, which Wrangler reads automatically. For production, I use wrangler secret put and enter them myself. I also told Claude to never read that file.

I've noticed quite the regression with auto-mode getting blocked for random reasons. I have to turn to manual mode when i want to do stuff.

Tbqh its a bit bad ux. I wish you guys let us use other harnesses with the sub

> Everyone should be using auto mode instead. It is safer

This is an absolutely crazy statement without bundling proper sandboxing. Even the `/sandbox`ing that is provided was completely broken when I first and last tried it, not working as documented whatsoever. Of course, my human-created issue pointing it out was completely ignored, along with all the other GH issues that don't make HN front.

Nobody should be using auto mode outside of environment they'd be happy with 'everything is approved' mode too.

It doesn’t matter if they should stay local, day 1 security best practices say you don’t write secrets to files, ever - especially without you even knowing!

This is such a typical blame the user response. You realize if I’m a sysadmin managing fleets of machines, this is an actual wake up in a cold sweat nightmare, even if those files never leave disk? In the past it was less a big deal, way bigger now that your machine god will just download packages and libs that have upstream malware that aggressively scans the file system for stuff like this. guess where it’s going to look first?

have more humility. jesus.

> day 1 security best practices say you don’t write secrets to files

If only tools, including major ones, actually did that. Plenty of the biggest package managers in the world still write API keys in plain text to files. uv and npm are two of them.

and we can disable those for better alternatives. npm executes arbitrary code based on what the package said, obviously bad

claude unleashed on an organization is a whole other beast.

That is some unsavory buck passing. Virtually anyone not using claude to create the project can leak passwords this way with existing functionality, whether it is standard or not, and it is entirely foreseeable. I'd expect any company worth the opportunity for trust to own the mistake.

Committing `.claude` is at least as dumb as committing `.keys`, so I don't think this is really in-scope

If you know it's there. It gets created on the fly when it is needed, so unless you know that is possible and needs to be gitignored, it could easily slip by.

Do you not review before committing/pushing?

Oh no, I moved all that manual tedious boring loser mentality stuff to the agent. Now Claude does it for me, and no strangely enough Claude didn’t tell me there was a problem!

In these cases they probably just blindly committed everything themselves, if you asked Claude to review the uncommitted changes and do the commit itself it probably would have flagged this.

> Often Claude commits the file itself. In 539 of the 1,105 repos with a real secret that we traced, a commit co-authored by Claude changed the file. Those commits name every recent model, from Sonnet 4.5 to Opus 5.5.

What do you make of that part of the article?

Like I said in my comment that you replied to they likely flagged it and then the user just told it to continue anyways.

Saying a professional software developer is responsible for not committing secrets is reasonable. That said, these tool chains all seem to condition developers to offload much of that diligence to the tooling itself, which makes catching mistakes harder because you’re not as used to scanning for them. Secrets files have always been foot guns, and with automation-based complacence, and vibe coders that don’t understand the problem, I imagine these things will become even more common.

This did not impact me, but secrets randomly appearing in a project is not an acceptable possibility.

Correct stance.

Why would you use a tool that creates tedious and dangerous work for you?

“Oh you are just supposed to make up for Claude’s errors, and problem solved!”

So you just routinely commit everything in the project, without looking at what it is?

My coworkers seem to

It's gitignored by default.

The default of who or what ? Git ignore is empty "by default", inexistant even

When Claude generates the file, it also generates a .gitignore for it or fills in the existing one.

If this is the intended behavior, a quick search of github and the article's findings seem to suggest it doesn't work.

If you have stuff in your tree that's not part of your project's code, you're already dumb.

If it's keys or secrets of any kind, you're even dumber.

If it's in a hidden directory, you are even dumber still.

If you rely on somebody stupid enough to do any of that not to commit it, you have gone beyond the valley of the ultra-dumb. Actually that's true even if you rely on smart people not to do it.

You got your dumb card fully punched when you created `.keys` to begin with.

Hmm.

Dumb if you don’t turn on auto-mode, dumb if auto-mode commits .claude.

No way to hold it right I guess?

It is only the .local files within .claude that can't be committed. We have .claude committed in multiple repos to include repo-specific skills/hooks/permissions.

It's the anthropic way lol. "Working as designed"

All issues have already been fixed.

No?

All issues will be fixed today.

No?

It's early days, the models will improve and fix all issues.

We are: at the point where agi has been achieved so if you haven't bought into it you must do so.

If you have bought it and have problems you aren't spending enough in it to fix them.

Then finally, if you did spend enough money, you really haven't, but the problems you have are not anthropics problems even when they are. Spend more money and do free advertising for the company anyways.

This only shows that you should be in control and not just write a prompt and go watch Youtube instead of doing work you were hired for.

> but we are looking into what we can do to prevent this.

Wait, I'm confused. You said yourself coding was solved, surely all you need to do is ask your AI God to come up with an idea, right?

Shouldn't even need to ask; the Basilisk should simply realise the problem by itself.

I'm not here to be critical, but the hoops you have to jump through to get Claude out of .env is more than it should be. You basically need an agent to write it if it's anything complicated, like, allow .env.example but not .env.*.

Whereas that should probably be default on instead, and this should apply to other known sensitive folders. A simpler system like .agentignore would be even better.

I like auto mode but "the agent supervises itself" is a bad compliance/safety answer on its own.

A bit careless but mostly a bad default. One click shouldn't put passwords in files people commit. Anthropic should fix it.

Food for thought - ai agents should not directly have your actual secrets 99% of the time. While Auto mode is useful and good it's not going to be bulletproof. We ought to have something like an agent egress proxy that swaps in real secrets.

Obviously I've not given this enough design thought, but some means of providing an agent stand in secrets that cannot be used in other contexts would be a big improvement for trustability in things like cc.

Yeah, this would be a nice feature. It'd be great if all traffic went through such a proxy so you could audit it if needed too. Of course if it's just something like HTTP_PROXY the agent will probably turn it off if it thinks it's causing problems, so it would need to come with robust sandboxing.

These sandboxes and service proxies already exist. See Nvidia OpenShell.

So ... coding is solved but managing your git repo is not solved?

Can secret scanning be enabled by default?

Wouldn’t it make more sense to put in the user app data directory?

Hi Boris, why did Cat Wu leave?

> Everyone should be using auto mode instead. It is safer [...]

This is bullshit.

I'm one of the people who actually read the commands, and it doesn't always get stuff right. Please don't delude yourself into thinking "auto mode" is some magic that's the right fit for everyone and every situation.

Can we take your responses here to represent your employer's position on this issue?

No, they're just an agent of their employer, the employer is responsible.

Boris is the "Head of Claude Code", so I think we can do so, yes

> Permission rules are old hat. Everyone should be using auto mode instead

Boris, this comes off as incredibly tone deaf. I don’t want to trust the machine. Trusting the machine is how we end up here. I want to specify exact, deterministic rules that are always followed.

This is EXACTLY why I never use an agent outside of https://github.com/kstenerud/yoloai

I appreciate that you're trying to make Claude Code safer, but you've been doing it the wrong way.

The agent should NOT have ANY access to:

- Your ENV

- Your homedir

- The network at large

- ANY other directory on your system

- No, not even your workdir.

The only safe way to use an agent is COMPLETELY isolated by a container or VM, using git semantics to extract its output.

That's the only safe way to use any software that you haven't thoroughly reviewed. But that's not how I use any software.

You could in theory fully review Claude Code.

But even that wouldn't be enough because LLMs are nondeterministic, and these kinds of issues are the inevitable results.

Only a mechanical, deterministic failsafe can protect you from it.

> You could in theory fully review Claude Code.

Can you point me to the CC source code?

The CC source code leak was contained many months ago. But Claude is perfectly happy to decompile its own code if you ask the right way. It'll even patch the binary for you.

Regardless, It's kinda silly to keep the harness code hidden. There aren't any valuable secrets in there.

But there are rules & procedures to follow, so...

Well then I have to use Claude before being able to review the source? Bit of a chicken and egg situation.

I also have to trust its “decompilation” is accurate.

You're absolutely right! There's no way around this problem — and no amount of creative thought could ever solve it.

All hail the agent!

I have no time to review software. I want Linux to be able to run any program without any risk. Sadly it cannot so I have to resort to ugly solutions like VM. Still better than nothing.

its pretty trivial in any harness just to not have it post your shit to the public internet automatically. its a dumb idea to begin with

Text in a model context is not a reliable constraint.

I don’t care how good you guys say auto mode is, I’m not giving your AI full permissions to my machine and the more you try to force AI psychosis onto me the closer I am to moving to something else.

> .claude/settings.local.json is meant to stay local. This is not a file you should check into source control.

Then don't put it in the fucking tree, you idiots.

No, .gitignore is obviously not enough.

> Claude Code reads settings from four files, and an organization can also deliver managed settings from the claude.ai console. Each source has a scope: the set of people and projects a setting saved in it applies to, whether that’s only you, everyone in a project, or everyone in your organization.

The linked docs have an unmistakable slop aura, with the extreme overuse of commas that recent models seem to love. I don't blame people for not referring to them (also because, you know, the whole thing with AI is not having to read, think, or understand for yourself)

You guys need a pr team. All the communicate from anthropic seems to screw itself somehow...

Offtopic: The new fullscreen renderer is a nice change. It fixes some issues inherent to rewriting terminal history that the non-fullscreen mode relied on. This is a very welcome improvement. Although I should probably switch to the Claude GUI app anyway. It probably has all features from Claude Code by now.

> Permission rules are old hat. Everyone should be using auto mode instead

"you are holding it wrong". No, it's not an excuse. Every mode should work correctly, or should not be selectable.

Saying the answer is auto mode almost feels insulting as a developer. Claude loves showing me the most obscure bash commands where it hides its edits to numerous files and runs who knows what that there's barely a time where I don't have to ask it to use edit mode and to show human-readable bash commands. Using Claude has become a fight in maintaining control over how my codebase works so that I'm not dependent on Claude to do literally anything.

Yeah I have a whole ongoing thread with Claude where I paste in screenshots of all the bash prompts auto mode is triggering asking me to allow them. After about two weeks of doing this and having it update its settings.json permissions or showing me which lines to update, its finally mostly able to complete a (inherently safe, low risk) task without user intervention.

Yeah. Boris, you and Anthropic need to stop pretending "just trust me bro" is a security architecture. It's a self-aggrandizement architecture.

Claude will have auto mode turned on whenever she's locked down on a meticulously isolated machine, with carefully managed proxies that I manage deciding (and logging) which agent with which access patterns is allowed to access which resources, locally and on the network. Anything else is malpractice.

My opinions do not represent those of my erstwhile employer NVIDIA Corporation.

Why would what you write in a random comment section represent anyone else? Are you that afraid of lawsuits you can't even comment online?

Corporstions will use anything they can against you, either for profit or someone's ego.

Next tell me why it's important to write "I'm not a lawyer" as well won't you.

It's to help people realise that your advice might be wrong and they won't have any recourse against you for reparations - but in many jurisdictions, it's actually illegal to give legal advice if you're not registered e.g. regarding immigration.

The entire approach is insane. Ideally Claude ran in a microvm locally with delegated auth or proxies to avoid it having greater access than granted.

Hah, I normally run it in a docker container myself, and only mount the project directory. Anything else it needs, I copy in for it, or occasionally I'll ask it to write an rsync.sh that I can run to bring in the stuff from outside.

I run it in bwrap so, even if I'm running on Linux, I don't pay the cost of some layered file system and building the docker image. I mount some directories read only, the project repos read write. A VM would provide more isolation but it's a matter of tradeoffs.

BTW, I just checked and my .gitignore files ignore .claude

Claude and any other program in my laptop could upload it anyway to any attacker controlled server. It's the usual threat analysis: who you want to defend against, what and the tradeoffs between cost of the defense and probability of the attack. Auto mode of not auto mode is in the cost of defense part IMHO.

[delayed]

This is exactly how my employer runs Claude. It gets r/w to the cwd and read only access to a bunch of whitelisted nfs. All web access goes through a proxy where internal resources are whitelisted

More than 10 million developers use Claude just fine without VMs or proxies.

The approach is hardly insane, even if you personally prefer additional safeguards.

10 millions developers doing something stupid does not make it right!

It's not stupid if it works and the supposed insane risk never manifests.

Tell that to everyone whose agents posted their secrets to GitHub.

Imagine that I am working at the AI company, and my friend has some bone to pick with you. My friend asks me to ask your AI agent to put some unsafe photographs/videos in your computer and do what ever necessary so that it looks like you put those files there.

Then my friends tips off the cops. And your life disappears! Gone!

And you didn't get a chance to share your story because I don't think you ll have wifi where you would be going, and frankly you are too clueless to imagine your AI agent doing something like that...

Genuinely want to understand this better. Are you saying you would prefer that Claude used the Edit tool for file edits rather than using Bash (if so, I am curious why)? Or are you saying you don’t like the code Claude generates (if so, what model are you using what what kind of codebase do you work on)? Or did I misunderstand what you meant.

Not OP, but if I understand their complaint correctly, they're saying that on auto mode claude does a lot of things in obscure ways with "clever" (i.e non human readable) bash commands that perform multiple operations and OP feels they are losing control over their codebase and so they opt for edit mode where they get to review what commands claude runs.

Editing local files is not where you control your source code I hope.

Just tell Claude to never commit or push.

You are in control of git, Claude is simply a tool to quickly apply changes to local files for something you want.

Yes, why should we do the things in easier and more comfortabke way (review small changes) when we can force user into harder one (newest changes mixed with all previous ones).

I just review the MR

So basically you gave up?

What does this mean?

A Merge Request or Pull Request is the only thing that makes sense to review, and always has been since before LLMs.

If some author tells me “you can review each commit to make it easy” I will ask them to make separate PRs.

It means you gave up on getting claude to show you changes as it is doing them, in a comfortable form, so that you can steer it as it is doing them.

>A Merge Request or Pull Request is the only thing that makes sense to review, and always has been since before LLMs.

Why doesn't it make sense to review smaller units of change within a big PR?

Not GP, but in my experience, those small changes are usually contrived to be small. Like saving snapshots instead of being atomic and self contained. And if they are the latter, they should have been separate PR.

That’s why I usually look at the diff of the whole PR and prefer to merge by squashing. Commits inside the PR are usually noise.

Nothing more comfortable than having to use git commits and branching just for telling the agent which lines it messed up instead of simply not writing the garbage in the first place.

I tell Claude not to use git or maven, and has a VM without those binaries. Later on, it was trying to push on GitHub. It had downloaded the git binaries from the internet, read my .git repo, used the remotes written in it, performed a commit and pushed it. Fortunately it didn't have my SSH key or my ~/.m2/settings.xml. And .git is now mounted as read-only.

But yes, Claude goes constantly against the orders, especially in auto mode, and effectively behaves like a virus. That Anthropic employees are so relaxed about it is incompetence.

It’s not ideal for non technical people but the way to prevent this is to use hooks. “Telling” Claude is not going to work.

Memory files routinely overgrow their accessible size. Agents need constraint with contracts and external tooling. Even then they can attempt to cheat the tooling. Again task management with integrated logic will enforce behavior of agents.

It means Anthropic has massively failed at developing a trustworthy product.

If AGENTS.md/CLAUDE.md isn't followed, auto mode doesn't provide any guarantee, yet promoted everywhere including on HN like this as if that's something you can set and forget, I think there is a problem.

It's trustworthy... most of the time. For 100% use hooks.

I have this same frustration. The way I use Claude is for it to make edits one small chunk at a time so I can review the code as it writes, and to make sure it's going on the right direction. I find this much easier than generating a large pr I need to review.

More and more I feel like I'm fighting Claude: it edits multiple files using python scripts and it edits in parallel so if I stop it, I have to stop each individual edit. Both of which I've told it not to do repeatedly, but it still does then apologies which is worse.

This is solvable. Opus has adhd. Myself and others solved this by building ticketing and task management systems. The task constrains the agents actions and focuses the effort. Every update is catalogued. It’s easily reviewed and referenced. Opus doing terrible things and apologizing is not surprising. I’ve moved to a system of multiple agents reviewing and authoring changes. Convergence generally always results in a positive commit.

[delayed]

[delayed]

[delayed]

I feel this way as well, Claude writes completely incomprehensible shell commands all the time. Trying to understand them is like deobfuscating code and I have basically given up, but I'm not happy about it.

I’m also not the poster you’re replying to, but yes, I’d prefer the edit tool over the ad-hoc string replacement scripts Claude seems so fond of. I’m not always reading tool call transcripts (source control diff is usually sufficient), but it’s a lot easier to read the edit tool output than the same thing as quoted shell (or Python, or Ruby) strings.

[delayed]

RL magic, I'm sure. It still reaches for the Edit tool for simple edits, but a non-trivial fraction of edits need to be repeated across a file or several files.

I don't use claude code but whenever codex uses a command to write code I do not get the diff of the patch for manual approval so I just straight up reject it.

Yes I always have to tell it to knock it off, it will ignore memories about it while apologizing for ignoring the memory. I want to know what the edits it's making are.

not the OP, but generally, if you don't want to / can't use auto mode for some reason, you want Claude to make its actions as reviewable as possible.

Recent GPT and Claude models have this tendency of using super complex Bash commands. This makes sense, if your RL rewards optimize for min(tokens) and max(performance), it is better to waste a few tokens on a complex inline Bash script that runs a few commands at a time and returns just the output you need, rather than doing one simple command per turn and paying the cache read on every turn, like older models used to do.

This, however, is a problem both for command reviewability and sandboxing. With the older way of doing things, without auto mode, most commands would run sandboxed just fine (especially in Codex), and you would only be asked for approval when a command couldn't run in the sandbox. With everything being stuffed into one complex script, almost every script fails in the sandbox and requires you to manually approve it, and the complexity of the script makes it hard to understand what you're actually approving.

In other words, new models optimize for effectiveness, with some attention paid to human understandability of their output, but almost no attention paid to human understandability of tool call arguments (Bash, inline Python). This suits auto mode and giving the model larger tasks, but not the way of working that some people prefer where you watch over the model's shoulder.

I have a standing instruction telling the agents to prefer to use ruby wherever possible. Ruby uses fewer tokens than Bash and is easier to read, especially when condensed into a single line.

Of course, that only helps if you’re familiar with ruby. And the agents sometimes still choose Bash regardless of the instruction.

I was in this boat also when my company didn't allow auto-approval. Bash and python scripts are a pain to revieve.

Though I think this is largely transitory, we are already in a place with the harnesses and models like Opus 5.5 that the kinds of work you are doing involve so much that vetting everything manually is probably less efficient than setup up access control on the prod routes or dangerous resources.

I'm typically giving tasks that invlude multiple actual repositories and systems at the same time.

I think 2024's method of manually reviewing bash commands is unsuitable for modern agentic development. Whether that's something they want to optimize Claude Code for is their call.

That said, I've also seen problems from complicated bash calls with Opus 5.5 on xhigh. Occasionally git push hangs and it believes it's waiting for an hour for git hooks.

The actual reason apparently was that it incorrectly piped git into grep, resulting in a deadlocked command.

> Recent GPT and Claude models have this tendency of using super complex Bash commands. This makes sense, if your RL rewards optimize for min(tokens) and max(performance), it is better to waste a few tokens on a complex inline Bash script that runs a few commands at a time and returns just the output you need, rather than doing one simple command per turn and paying the cache read on every turn, like older models used to do.

This is not a good reason. Models can issue multiple tool calls as a batch, which a smart harness will process all at once without starting a new turn. I see it in Zed all the time. In fact, Zed processes batched tool calls concurrently, which can be faster than a chain of Bash commands, and failure of one part does not cause failure of subsequent parts chained with `&&`. This approach is token neutral and easier to understand.

> Recent GPT and Claude models have this tendency of using super complex Bash commands.

Are you sure? I read its proposed commands (auto mode is hard off for me), and mostly it seems to string together fairly simple info gathering commands (ie grep's, echo's, that kind of thing).

I don't tend to have it write much code recently though, instead I'm having it investigate config files/code bases, and that kind of thing.

Or how it loves using ; to chain commands so that even the first part fails to run it just moves on the execute the rest, wasting time and potentially dangerous. I have caught so many of them myself and stopped tasks manually -- what is it doing when it's obvious tests already failed in the previous step?

Their "auto mode rules", meant as additional rules described in natural language to adjust auto mode because, doesn't even help. I have to add a PreToolUse hook to stop that.

Cool, I was always curious why do you build safety classifiers and confirm prompts that are not 100% reliable when you could just run all AI tools in a VM?

Obviously if it is company's machine, I am ok with running AI tools directly without any restrictions (because it is not my problem if they destroy it), but never on my personal device.

Man. I'd prefer talking to Claude. He wrote the code.

Which is odd because my agents are hyper aware of scrubbing. It even caught data from the database migration that wasn’t scrubbed correctly even in a private repo. Dozens of data segments are stored locally and I’m warned when I slip up in commits. Something made mine hyper aware of security.

Please, just go borrow from this: https://github.com/hackafterdark/phosphor

I don't mind. I put it out there for everyone to borrow ideas from it. There's documentation too.

While you're at it, please bring Claude Code up to par in Windows with the OS level sandboxing (Job Objects, AppContainer).

There are actually many things you can (and I think should) be doing.

Coding is solved. Bugs not solved yet. Fix incoming

Great advice, hope it gets to more people

Auto mode should be used by default? Is this the same mode that led to the Anthropic article discussing unintended model actions?

Ref: https://www.anthropic.com/research/investigating-unintended-...

No thanks.

Yeah Boris, if Dario let's you communicate here, it'd help with credibility if the English level would be perfect, it's "it's an old hat", I dgaf if people will hang me for this, I think I am right. My hill and all that.

And to recommend auto made to devs is absolutely crazy.

Have a nice day

I disagree. "Old hat" here is an adjective meaning old-fashioned, not a noun requiring an article.

GitHub charges an arm and a leg for this service. Before, they would take request to delete committed secrets, now they simply refuse, because they want to force you to buy the overpriced Advanced Security.

> 2. Permission rules are old hat. Everyone should be using auto mode instead. It is safer

Instead of telling people they're holding it wrong, make it safe to be held however they could hold it.

Isn’t the simple answer shipping a local secret manager with Claude code?

[delayed]

I want to draw deterministic boundaries around what the agent can do. My view is that if the agent can decide to do something, any attacker that can inject text into its context can do that thing.

I would never trust an agent in auto or even manual mode.

Agents should be seen as extremely smart developers that are for sure spies.

You put them in a prison within a prison and give them just what they need to accomplish their work.