Claude Code now supports hooks

This discussion centers around the introduction and adoption of "hooks" in Claude Code, a feature that allows for more programmatic control and runtime verification of an AI coding agent's behavior. The conversation spans user enthusiasm, comparisons to other tools, practical implementation details, and concerns about the terms of service and the future economic viability of such services.

Enthusiasm for Hooks and Enhanced Control

Many users expressed excitement about the "hooks" feature, viewing it as a significant advancement that enhances the capabilities and reliability of AI coding agents. The core benefit highlighted is the ability to implement custom logic and validation at specific points in the agent's workflow, moving beyond simple instruction following.

• "Really excited to see this implemented. Hooks will be important for 'context engineering' and runtime verification of an agent's performance." - termed ramoz
• "This is a pretty killer feature that I would expect to find in all the coding agents soon." - termed chisleu
• "This closes a big feature gap." - termed brynary
• "This is great, it means you can set up complex concrete rules about commands CC is allowed to run (and with what arguments), rather than trying to coax these via CLAUDE.md." - termed nojs
• "It looks like the events that can be hooked are somewhat limited to start, and I wonder if they will make it easy to hook Git commit and Git push." - termed brynary
• "I would like to limit it to doing certain essential aspects of workflow with the commands I specify, just as a developer would normally do. Are 'Hooks' the right answer?" - termed jonstewart

Comparison and Competition with Other Tools

Several users drew comparisons between Claude Code and other AI coding tools, particularly Cursor, and discussed the competitive landscape. The discussion touched on feature parity, proprietary aspects versus open-source alternatives, and the overall value proposition of different platforms.

• "I really wish they’d just let me fork the damn thing already." - termed wilde (implying a desire for a more open-source or customizable solution)
• "Would love to see this in Cursor. My workaround right now is using a bunch of rules that sort of work some of the time." - termed petethepig
• "Cursor is still the best when you don’t have access to a Claude subscription." - termed Aeolun
• "That tab autocomplete and predicting what I'm going to edit next is the best I've found. The rest I can take or leave (plenty of good or better alternatives)" - termed porker (referring to Cursor's features)
• "Not to take away anything here, but hooks are present in other similar products. Atleast one example here - [link to AWS Q developer user guide]" - termed wippler
• "Amazing how there's whole companies dedicated to this and yet claude code keeps leading the way. " - termed cheriot

Practical Implementation and Limitations

Users discussed the practicalities of using hooks, including how they integrate with existing workflows, potential debugging challenges, and the need for more granular control. The conversation also highlighted instances where CLAUDE.md instructions might be missed by the AI, underscoring the value of a more deterministic hook system.

• "you can just tell it do that or in your claude.md. don't need hooks" - termed apwell23
• "In some cases cc misses rules in CLAUDE.md." - termed thelittleone
• "In many cases." - termed ramoz (agreeing that CLAUDE.md can be unreliable)
• "so its back to hand coding stuff again. They are going to slowly add 'features' that brings handcoding back till its like 100% handcoding again." - termed apwell23 (expressing a concern about feature creep leading back to manual coding)
• "This doesn’t come without consideration. You can see I mention this in the original feature request. Yes - it’s fine to think of it as handholding (or handcoding). These model providers cannot be responsible for ultimate alignment with their users. Today, they can at best enable integration so a user, or business, can express and ensure their own alignment at runtime." - termed ramoz (defending the need for hooks as a mechanism for user-defined alignment)
• "1) Assign coding task via prompt 2) Hook: Write test for prompt proves 3) Write code 4) Hook: Test code 5) Code passes -> Commit 6) Else go to 3." - termed thelittleone (outlining a potential workflow using hooks)
• "Without Opus (larger model), for $20 you get only Sonnet. $100 and $200 plans have Opus." - termed vl (discussing pricing tiers and model availability)
• "This also: [pre] Exit Code 2 Behavior PreToolUse - Blocks the tool call, shows error to Claude [/pre] This is great, it means you can set up complex concrete rules about commands CC is allowed to run (and with what arguments), rather than trying to coax these via CLAUDE.md." - termed nojs (explaining a use case for hooks in command execution control)
• "You can already do this in .Claude/settings.json" - termed rco8786 (pointing out an alternative configuration method)
• "Ah you’re right, but for more complex logic it’s useful to be able to run it through a custom script" - termed nojs (clarifying the need for hooks over static settings)
• "We’ve been using CLAUDE.md instructions to tell Claude to auto-format code with the Qlty CLI (...) but Claude a bit hit and miss in following them. The determinism here is a win." - termed brynary
• "This needs a way to match directories for changes in monorepos. E.g. run this linter only if there were changes in this directory." - termed parham
• "Despite explicit directions in CLAUDE.md to build with 'make -j8' and run unit tests with 'make -j8 check', I see it sometimes running make without -j or calling the test executable directly." - termed jonstewart (experiencing a lack of determinism even with CLAUDE.md)

Concerns Regarding Terms of Service and Legal Ambiguities

A significant portion of the discussion revolved around Anthropic's terms of service (ToS) and the perceived ambiguities and potential restrictions on user activities. Users questioned what constitutes "competing" with Anthropic and whether common development tasks might inadvertently violate these terms, leading to a discussion about the practical implications and the "trust and safety" aspects.

• "Given the Anthropic legal terms forbid competing with them, what are we actually allowed to do with this? Seems confusing what is allowed." - termed bionhoward
• "No machine learning work? That would compete. No writing stuff I would train AI on. Except I own the stuff it writes, but I can’t use it. Can we build websites with it? What websites don’t compete with Anthropic?" - termed bionhoward (listing potential ambiguities)
• "Feels like the dirty secret of AI services is, every possible use case violates the terms, and we just have to accept we’re using something their legal team told us not to use? How is that logically consistent?" - termed bionhoward
• "Would you argue that Cursor (valued at $10B) is breaking Anthropic's terms by making an IDE that competes with their Canvas feature?" - termed nerdsniper
• "Cursor isn't building models trained with the outputs of Anthropic models (I think). That's what the ToS is forbidding." - termed varenc (offering an interpretation of the ToS)
• "Oh come on, your CRUD app is not competing with an LLMaaS" - termed jazzyjackson (challenging a strict interpretation)
• "You’re only competing with them if you’re doing something they consider competitive. OpenAI is competitive, you are not" - termed paulsutter
• "Anthropic's terms typically restrict training competing AI models with their outputs, not building standard applications or websites that simply use their API as a tool." - termed ethan_smith (providing a clearer definition of the ToS restriction)

Cost and Value Proposition of AI Services

The economic aspect of using AI coding tools was also a recurring theme. Users discussed subscription costs, usage-based billing, and whether the value provided justifies the expense, particularly in the context of alternative tools and the long-term sustainability of these services.

• "To be frank? I can't justify paying for a single-purpose LLM service subscription: Cursor has have a 1-year free educational plan, and for general-purpose multimodal reasoning model work (e.g. OCR, general knowledge reference, math computations, prose processing), I already have a ChatGPT Plus subscription. It's the streaming service dilemma all over again." - termed nxobject
• "Claude Code has basically grown to dominate my initial coding workflow. I was using the API and passed $50 easily, so I upgraded to the $100 a month plan and have already reached $100 in usage." - termed bearjaws
• "It remains to be seen whether it's a net value once the VC firehose dries up and the true costs are revealed. It's quite possible that the profitable price is not worth it for most companies." - termed billbrown
• "Opus-4 feels like what OAI was trying to hype up for the better part of 6 months before releasing 4.5" - termed bearjaws
• "Just started using Claude (very late to the game), and I am truly blown away. Instead of struggling for hours trying to get the right syntax for a Powershell script or to convert Python to Go, I simply ask Claude to make it happen. This helps me focus on content creation instead of the mind-bending experience of syntax across various languages." - termed rtp4me
• "At this point, my goal is to get good at asking the right kinds of questions to get the best/most accurate answers." - termed rtp4me

The Nature of AI and the Need for Human Oversight

A philosophical undercurrent in the discussion touched on the fundamental nature of LLMs, comparing them to infallible systems versus more fallible tools requiring careful management. The need for human intelligence in configuring, verifying, and ultimately directing AI remains a critical element.

• "So many people yearn for LLM's to be like the Star Trek ship computer, which when asked a question unconditionally provides a response relevant and correct, needing no verification. A better analogy is LLM's are closer to the 'universal translator' with an occasional interaction similar to [Monty Python Black Knight quote]" - termed AdieuToLogic (drawing an analogy to highlight the fallibility and potential for miscommunication with LLMs)
• "The nature of these systems already requires human symbiosis. This is nothing more than a new integration point. Will empower agents beyond today’s capabilities, increase adoption." - termed ramoz
• "Claude Code has basically grown to dominate my initial coding workflow. (...) Still have to audit every change, commit often, but it works great 90% of the time." - termed bearjaws
• "As an aside, people say AI will eliminate coding jobs, but then who will configure these hooks? Or think about adding such a feature? These kinds of tooling and related work will still be there unless AI evolves to the point that it even thinks of this and announces this to all other AI entities and they also implement it properly etc." - termed mkagenius (suggesting that AI development itself creates new needs for skilled human labor)