I think most skills should have disable-model-invocation enabled by default—they’re far too easy to trigger accidentally. In other words, skills should default to disabling model invocation unless you explicitly opt in.
@colemurray@claudeai Big OpenInspect fan here, but the Slack integration is on a much higher level than OpenInspect's. It's worth a look to draw some inspiration :)
@0xSero@moskstraum21745 Another thing that works great, especially when you’re starting from scratch, is asking to have it explained in increasing levels of complexity
It’s basically running git blame on your thoughts instead of on the code: your reasoning flaws are exposed publicly, but it also becomes an extraordinary feedback loop.
Just checked out @EntireHQ. I like that you have more context about who wrote what, token usage, prompts. No full Cursor support yet, hoping it will come soon. But worried about, that you can then see all my "wtf you destroyed my whole page layout" prompts publicly forever in git
@swyx@ankitxg Whatever level humans intervene in code review, call it the final boss is wrong. There are other steps that remain a bottleneck in the production or execution of applications.
@colemurray@Ryan_And3rs0n nice to hear scheduled tasks landing soon :)
would be great also to ease the creation of different agents according to task type/use cases. I like how Oz do it with skills
@colemurray What should I do if I need other dependencies for my modal image? Do I necessarily have to fork the project and modify .openinspect/setup.sh or modal-infra/src/images/base.py?
@gmickel A pattern that also works well is when the developer shares the conversation/session they went through up to the PR. This makes it possible to identify gaps in the reasoning, or sometimes issues with the initial framing that wasn’t quite right.
Ideal for searching a specific session when constantly juggling between many agents and you no longer remember which one you created it with
github.com/angristan/fast…
People wondering why some devs prefer CLI agent other IDE. If there was just one reason, watch this video: youtube.com/watch?v=BsAHun… Lots of cool stuff yes but where does it mainly happen? In 2 strips of 100px wide. Not much in the largest central area.
gritql + husky > agents․md
models can't read your entire codebase to understand architecture patterns, but you CAN enforce this at commit time with linters instead of wasting tokens filling up non-deterministic rules files.
we use a husky pre-commit hook that runs our linter
@badlogicgames Interesting. I need to check what other agents harness do in their system prompts to prevent that behavior. I'm only see it in pi. I'll send you a session sample in DM if you don't mind. I would be interested to know if that's the kind of behavior you usually see.
@badlogicgames Are there any settings or commands that would make Pi generate fewer documents? It tends to create a large number of documents and recap even for small operations.
322 Followers 526 FollowingLexicon Labs Publishing is dedicated to merging learning with fun, offering a diverse array of educational materials for all ages.
594 Followers 285 FollowingHead of Training & Inference @LightOnIO | PhD in physics of the universe 💫.
Interested in IR, Evals & synthetic data.
I also like gaming 🎮 and lifting 🏋️♀️.
1K Followers 2K FollowingContinuous delivery simplified on Kubernetes
Automate your code to test to production across environments, and apps in few steps.
4K Followers 4K FollowingTechFAIR LIVE is the Community Tech Festival, with 10,000+ tech professionals, enthusiasts and business leaders coming to learn and grow together. Join us.
743 Followers 357 FollowingConsultant, Contractor, Software developer, Speaker, Writer, Musician
Blog: https://t.co/pbrinxpgXZ
Definition of done: Working software, in production
7K Followers 5K FollowingSharing content in software development, testing, and AI to all the community. For more of our own insights, visit: https://t.co/e2CvX7Dsjm
11K Followers 102 FollowingCEO @Glowforge and https://t.co/7QcGGhWs9L, Wharton research fellow in AI. He/him. FormerGoogle, Sparkbuy, Ontela. Author, The Startup CEO Guidebook. Lucky dad.
23K Followers 2K Followingbuilding the post-IDE IDE at @humanlayer_dev, YC alum, @aitinkerers sf lead, prev @replicatedhq @SproutSocial @nasa - ai that works pod @ https://t.co/69BhaNtWfd
544K Followers 24 FollowingPersonal AI that actually does things. Your agent, your machine, your rules. Now a 501(c)(3) non-profit foundation — open and independent, forever. 🦞
22K Followers 4K FollowingMember of Technical Staff, AI Research @sullyai
Building AI medical employees
Prev/ AI Persona Development Research @ 99Ravens
2K Followers 700 FollowingLeading AI @ GrowthFactors (Bregal fund). Building, https://t.co/Ipgit3HXK7, https://t.co/2zCTYmSHkm, https://t.co/sSvNpINJKn, DocIQ.. Was orchestrating before it was cool.
26K Followers 26 FollowingWe advance the development of ASI and foster open source collaboration towards a smarter future.
Discord: https://t.co/BtsFsAUsvT
56K Followers 1K FollowingArmin's handler at https://t.co/B05ybKGkzx. Old man yelling at Claudes.
https://t.co/Q1wG57v1yc
https://t.co/mnOoWUr0TO
https://t.co/8i5vIRE0Wn
1.7M Followers 2 FollowingClaude is an AI assistant built by @anthropicai to be safe, accurate, and secure. Talk to Claude on https://t.co/ZhTwG8dz3D or download the app.
14K Followers 1 FollowingBuild & use computer vision models fast ✨
Get started: https://t.co/QvAsWtf7c8
Open source datasets & models https://t.co/IzwgoIkBEe
594 Followers 285 FollowingHead of Training & Inference @LightOnIO | PhD in physics of the universe 💫.
Interested in IR, Evals & synthetic data.
I also like gaming 🎮 and lifting 🏋️♀️.
47K Followers 73 FollowingCommunity account for sharing ClaudeCode related projects and releases. Views/shares independent from @AnthropicAI positions.
9K Followers 906 FollowingAI / Technical Artist / ComfyUI expert • Training models • Creating crazy workflow • Creative Engineer at https://t.co/SpUceQW1mb • SF based • Old Gobelins Paris