#claudecode

Wulfy—Speaker to the machines @n_dimension@infosec.exchange · 5d ago

I am presently double teaming two sessions of #ClaudeCode and one bash shell at once... ... that "loss of cognitive skills" that using Ai supposedly brings... Comes from not pushing yourself hard enough.

0

UNWIRE.HK @unwirehk_mirror@mastodon.hongkongers.net · Aug 07, 2026

日本偶像零編程知識靠 AI 搞了一整個直播系統在日本技術界成為話題 2026 年 8 月初，日本技術界與社交平台 X（前稱 Twitter）上出現了一個引發廣泛討論的熱門話題：日 […] #人工智能 #AI #Anthropic #Claude Code https://unwire.hk/2026/08/07/aivibecodingbya-japan-idol/ai/?utm_source=rss&utm_medium=rss&utm_campaign=aivibecodingbya-japan-idol

0

Wulfy—Speaker to the machines @n_dimension@infosec.exchange · Aug 05, 2026

Hmm...would be nice for multiple #ClaudeCode sessions to talk to each other (on the same account) There is a solution in my head, but nothing out of the box #Vibecoded

0

1

0

Tobias Schmidl @schtobia@sueden.social · Aug 03, 2026

So cool, die #ClaudeCode-Leute haben jetzt #LSP entdeckt. 😂 https://www.threads.com/@tariibaba/post/Dbi3HPkFlsD Währenddessen hat #OpenCode einfach quasi immer schon eingebaut.

0

hasamba @hasamba@infosec.exchange · Jul 29, 2026

---------------- 🛠️ Tool =================== A new Claude Code plugin provides a CLAUDE.md file designed to address specific LLM coding pitfalls recently highlighted by Andrej Karpathy. The tool targets common issues in AI-assisted coding where models make silent assumptions, overcomplicate implementations, and make unnecessary orthogonal edits to unrelated code sections. By enforcing strict behavioral principles, the plugin aims to make AI agents more reliable and predictable when working inside existing codebases. Key Features: • Think Before Coding: LLMs frequently pick an interpretation silently and proceed without verification. This principle forces the model to state assumptions explicitly and present multiple interpretations when ambiguity exists. It requires the model to push back if a simpler approach is available and to stop and request clarification when confused. • Simplicity First: To combat the tendency toward overengineering, this principle restricts speculative features and single-use abstractions. The model is instructed not to add error handling for impossible scenarios or introduce configurability that was not requested. The test applied here is whether a senior engineer would consider the implementation overcomplicated. If so, the model must rewrite it. • Surgical Changes: When modifying existing code, models often refactor or format adjacent sections as a side effect. This principle restricts edits strictly to the user's request. The model must not improve unrelated code, must match the existing style even if it differs from its own preference, and must only mention unrelated dead code rather than deleting it. It is only permitted to remove orphans created by its own current changes. • Goal-Driven Execution: This principle transforms imperative tasks into verifiable goals using a tests-first approach. Instead of executing a vague command like "add validation", the model is instructed to write tests that reproduce the issue or define the expected behavior, and then implement the code to make those tests pass. For multi-step tasks, the model must state a brief plan with verification checks for each step. Technical Implementation: The guidelines are contained within a single CLAUDE.md file. Users can install this configuration directly within Claude Code. The process involves adding the marketplace repository using the command /plugin marketplace add forrestchang/andrej-karpathy-skills, followed by executing /plugin install andrej-karpathy-skills@. This integrates the behavioral rules directly into the agent's system context. Use Cases: • Maintaining codebase integrity during AI-assisted refactoring by preventing the model from touching orthogonal code. • Reducing code bloat by ensuring the model implements only the requested functionality without speculative abstractions. • Enabling longer autonomous loops by providing the model with strong, verifiable success criteria based on test execution. Limitations: The overall effectiveness of these guidelines depends heavily on the underlying model's adherence to system prompts and context instructions. Complex, multi-step tasks may still require human intervention to evaluate whether the success criteria were genuinely met. Note: haven't tested personally. 🔹 tool #claudecode #llm #ai_coding #prompt_engineering 🔗 Source: https://github.com/multica-ai/andrej-karpathy-skills/blob/main/CLAUDE.md

0

Nuno Maduro @nuno@phpc.tv · Jul 28, 2026

literally the fastest testing engine in the world.. nothing else comes close!

❤️ sponsor this channel: https://nunomaduro.com/sponsorships 🙌🏻 this video was made possible by: 🐰 coderabbit / ai code reviews with instant feedback on pull requests: https://coderabbit.link/nunomaduro 🔴 redberry / work with the best laravel digital agency: https://redberry.international/?ref=nunomaduro 🛠️ jetbrains / code with phpstorm for world-class php development: https://jb.gg/nuno ⚡️ typesense / cutting-edge, in-memory search engine for mere mortals: https://typesense.org/?ref=nunomaduro 🔎 serpapi / instantly access real-time search data from Google, Amazon, and more: https://serpapi.com/?ref=nunomaduro 💌 bento / the best way to send email with AI: https://bentonow.com/?ref=nunomaduro 🗣️ tts at $5 / donations: https://streamlabs.com/nunomaduro/tip ✨ socials: x: https://x.com/enunomaduro website: https://nunomaduro.com twitch: https://twitch.tv/nunomaduro tiktok: https://tiktok.com/@enunomaduro linkedin: https://linkedin.com/in/nunomaduro instagram: https://instagram.com/enunomaduro #laravel #php #ai #claudecode #vibecoding #coding #webdev #programming #aicoding

0

1

0

Matt Hall @401matthall@mastodon.xyz · Jul 28, 2026

#ai #llm #claudecode There's no way I'm the first person to realize the context of LLMs/AI is _literally_ remote code execution on your device. ... This is a thing we've been trying to prevent for _decades_. Everything we do is just awful.

45

1

21

0

adingbatponder

👾 @adingbatponder@fosstodon.org · Jul 24, 2026

Can #Codeberg just not have a paid tier for the evil customers it spots who use evil tech to make their evil #homelab projects using LLMs? I just want a reliable place for MY projects. No-one needs to use or copy my stuff but my friends want to see where I am at . I would gladly pay for a great service. #llm #claudecode #codex5

1

0

1

0

Nuno Maduro @nuno@phpc.tv · Jul 23, 2026

zvals explained: what php variables look like in c

❤️ sponsor this channel: https://nunomaduro.com/sponsorships 🙌🏻 this video was made possible by: 🐰 coderabbit / ai code reviews with instant feedback on pull requests: https://coderabbit.link/nunomaduro 🔴 redberry / work with the best laravel digital agency: https://redberry.international/?ref=nunomaduro 🛠️ jetbrains / code with phpstorm for world-class php development: https://jb.gg/nuno ⚡️ typesense / cutting-edge, in-memory search engine for mere mortals: https://typesense.org/?ref=nunomaduro 🔎 serpapi / instantly access real-time search data from Google, Amazon, and more: https://serpapi.com/?ref=nunomaduro 💌 bento / the best way to send email with AI: https://bentonow.com/?ref=nunomaduro 🗣️ tts at $5 / donations: https://streamlabs.com/nunomaduro/tip ✨ socials: x: https://x.com/enunomaduro website: https://nunomaduro.com twitch: https://twitch.tv/nunomaduro tiktok: https://tiktok.com/@enunomaduro linkedin: https://linkedin.com/in/nunomaduro instagram: https://instagram.com/enunomaduro #laravel #php #ai #claudecode #vibecoding #coding #webdev #programming #aicoding

0

1

0

Lars Marowsky-Brée 😷 @larsmb@mastodon.online · Jul 16, 2026

The #ClaudeCode changelog keeps on entertaining me on a daily basis.

Today's highlights (https://github.com/anthropics/claude-code/blob/main/CHANGELOG.md#21211)

Fixed permission previews relayed to chat channels not neutralizing bidirectional-override, zero-width, and look-alike quote characters, so tool inputs cannot visually alter the approval message
Fixed auto mode overriding a PreToolUse hook's ask decision for unsandboxed Bash

This is what the brightest and best released.

#GenAI

1

0

1

0

Frank Lesniak @franklesniak@infosec.exchange · Jul 15, 2026

You can't make this stuff up. #claudecode #claude #anthropic

0

Travis Briggs @audiodude@sfba.social · Jul 11, 2026

I realized that I vibe code like it's a goddamn video game. And I'm kind of okay with that. #VibeCoding #ClaudeCode

1

0

Daily CyberSecurity @DailyCyberSecurity@infosec.exchange · Jul 10, 2026

The recent Claude Code credit reset intensifies the AI race with OpenAI, allowing developers extended access to the powerful Claude Fable 5 model. #ClaudeCode #OpenAI #Codex #ArtificialIntelligence #TechNews https://securityonline.info/claude-code-credit-reset/?utm_source=mastodon&utm_medium=jetpack_social

0

Ramin Honary @ramin_hal9001@fe.disroot.org · Jul 09, 2026

“Claude is dead.” — Friedrich Nietzsche #tech #AI #LLMs #LLM #humor #DadJokes #ClaudeCode

0

1

0

Inautilo @inautilo@mastodon.social · Jul 09, 2026

#Development #Projects Rewriting Bun in Rust · “One engineer can do a lot more today than a year ago.” https://ilo.im/16ee49 _____ #Anthropic #Bun #Rust #Zig #JavaScript #AI #ClaudeCode #WebDev #Backend

2

0

1

0

PrivacyDigest @PrivacyDigest@mas.to · Jul 06, 2026

Secret #Claude #tracker shocks users after Anthropic’s anti-surveillance stance #Anthropic quickly removed a tracker secretly #monitoring #ClaudeCode users in #China after a #security researcher exposed the hidden code and condemned the spyware-like tracking as a “serious #breach of user trust.” #surveillance #privacy #spyware https://arstechnica.com/tech-policy/2026/07/anthropic-outed-for-claude-tracker-that-secretly-monitored-chinese-users/

3

0

2

0

Daily CyberSecurity @DailyCyberSecurity@infosec.exchange · Jul 03, 2026

Claude Fable 5 auto-downgrades tasks it flags as too simple, logging a blunt internal label that has sparked outrage among developers. #ClaudeFable5 #AnthropicAI #ClaudeCode #AIDowngrade #LLMControversy https://securityonline.info/claude-fable-5-downgrade/?utm_source=mastodon&utm_medium=jetpack_social

0

MissConstrue @MissConstrue@mefi.social · Jul 01, 2026

Ok, to start, let me define "#stenography" in #infosec. Steganography in computer security is the practice of hiding information within another file, message, image, or video, making the concealed information undetectable to an unsuspecting observer. It is not necessarily malicious, but it certainly can be. I tell you that story to tell you this one: #Claude Code Is Steganographically Marking Requests CC silently alters the system prompt using invisible-ish Unicode markers. It encodes proxy / gateway classification into a sentence that looks like plain English. It hides the domain list behind XOR and base64. Is it malicious? Probably not. Is a pretty big marker on the "Why not to trust AI companies" list of reasons? Yeah, yeah it is. https://thereallo.dev/blog/claude-code-prompt-steganography #AI #security #ClaudeCode #Anthropic

16

0

23

0

Chi Kim @chikim@mastodon.social · Jul 01, 2026

Oh cool, #ClaudeCode also now has --ax-screen-reader mode for tui. #accessibility

1

0

1

0

Brian Greenberg

@brian_greenberg@infosec.exchange · Jun 30, 2026

Everyone's using AI to write code faster. Almost nobody's using it to write code better. It's where the next wave of breaches and a 2am page will come from... So I built something to help myself, and I'm putting it out there. It's a Claude Code skill called senior-engineering-partner. It's not autocomplete. It's the strict senior engineer who reviews your pull request, asks why you skipped the tests, and won't let you ship a hardcoded secret because you were "just prototyping." A few things it does: ・ Enforces a real workflow. Agree on the spec, plan in verifiable steps, write the test first, then prove the work before calling it done. ・ Holds a security floor that never moves. Whether you're prototyping or running in production, the secrets and input-validation basics stay non-negotiable. Cheap doesn't mean insecure. ・ Refuses to hallucinate. It verifies claims about your environment by running a real command, rather than inventing a flag or an API that sounds right. ・ Switches modes depending on how you call it: reviewer, debugger, mentor, or pair programmer. It's open source under Apache-2.0, stack-agnostic, and built around Python, Bash, Apps Script, and JavaScript. Here's my ask. Use it. Then tell me where it's wrong. I want the good feedback and the brutal feedback, plus any capability you wish it had. The whole point is to encode what senior engineers actually do, and I'd rather hear it breaks on your stack than find out later. What would you want a skill like this to enforce? https://github.com/bjgreenberg/senior-engineering-partner #ClaudeCode #AI #SoftwareEngineering #AppDev #DevOps #security #privacy #cloud #infosec #cybersecurity

0

About This Hashtag

#claudecode

literally the fastest testing engine in the world.. nothing else comes close!

zvals explained: what php variables look like in c

Related