| Server IP : 3.147.158.171 / Your IP : 216.73.216.216 Web Server : Apache/2.4.67 (Amazon Linux) OpenSSL/3.5.5 System : Linux ip-172-31-2-178.us-east-2.compute.internal 6.1.172-216.329.amzn2023.x86_64 #1 SMP PREEMPT_DYNAMIC Wed May 20 06:31:34 UTC 2026 x86_64 User : ec2-user ( 1000) PHP Version : 8.4.21 Disable Function : NONE MySQL : OFF | cURL : ON | WGET : ON | Perl : ON | Python : OFF | Sudo : ON | Pkexec : OFF Directory : /tsai/repo/api/logs/articles/ |
Upload File : |
# Tavily Search Results: Geneva **Query:** AI coding agents updates — Claude Code, OpenAI Codex, Cursor, GitHub Copilot, Gemini Code Assist **Topic:** general **Results:** 5 **Date:** 2026-05-04 00:06:56 --- ## 1. Claude Code, Copilot and Codex all got hacked. Every attacker ... **URL:** [https://venturebeat.com/security/six-exploits-broke-ai-coding-agents-iam-never-saw-them](https://venturebeat.com/security/six-exploits-broke-ai-coding-agents-iam-never-saw-them) **Score:** 0.7863 ### Summary Riemer drew the operational line in an exclusive VentureBeat interview. "It becomes, I don't know you until I validate you." The branch name talked to the shell before validation. The GitHub issue talked to Copilot before anyone read it. ## Security director action plan 1. Inventory every AI coding agent (CIEM). Codex, Claude Code, Copilot, Cursor, Gemini Code Assist, Windsurf. List the credentials and OAuth scopes each received at setup. If your CMDB has no category for AI agent identities, create one. 2. Audit OAuth scopes and patch levels. Upgrade Claude Code to 2.1.90 or later. Verify Copilot's August 2025 patch. Migrate Vertex AI to the bring-your-own-service-account model. [...] Audit AI-generated code for security flaws Anthropic launched Claude Code Security (Feb 2026). OpenAI launched Codex Security (March 2026).Both scan generated code. Neither scans the agent’s own execution environment or credential handling.Code-output security is not agent-runtime security. The agent itself is the attack surface. ### Full Content # Claude Code, Copilot and Codex all got hacked. Every attacker went for the credential, not the model. | VentureBeat [](https://venturebeat.com/) * [Orchestration](https://venturebeat.com/category/orchestration) * [Infrastructure](https://venturebeat.com/category/infrastructure) * [Data](https://venturebeat.com/category/data) * [Security](https://venturebeat.com/category/security) More [Newsletters](https://venturebeat.com/newsletters) # Claude Code, Copilot and Codex all got hacked. Every attacker went for the credential, not the model. [Louis Columbus](https://venturebeat.com/author/louis-columbus) 9:30 am, PT, April 30, 2026  VentureBeat created with Imagen [](https://www.google.com/preferences/source?q=venturebeat.com "Add to Google Preferred Source") On March 30, [BeyondTrust](https://www.beyondtrust.com/blog/entry/openai-codex-command-injection-vulnerability-github-token) proved that a crafted GitHub branch name could steal Codex’s OAuth token in cleartext. OpenAI classified it Critical P1. Two days later, Anthropic’s Claude Code source code [spilled onto the public npm registry](https://venturebeat.com/technology/claude-codes-source-code-appears-to-have-leaked-heres-what-we-know), and within hours, [Adversa](https://adversa.ai/blog/claude-code-security-bypass-deny-rules-disabled/) found Claude Code silently ignored its own deny rules once a command exceeded 50 subcommands. These were not isolated bugs. They were the latest in a nine-month run: six research teams disclosed exploits against Codex, Claude Code, Copilot, and Vertex AI, and every exploit followed the same pattern. An AI coding agent held a credential, executed an action, and authenticated to a production system without a human session anchoring the request. The attack surface was first demonstrated at Black Hat USA 2025, when Zenity CTO [Michael Bargury hijacked](https://labs.zenity.io/p/hsc25) ChatGPT, Microsoft Copilot Studio, Google Gemini, Salesforce Einstein and Cursor with Jira MCP on stage with zero clicks. Nine months later, those credentials are what attackers reached. Merritt Baer, CSO at Enkrypt AI and former Deputy CISO at AWS, named the failure in an exclusive VentureBeat interview. “Enterprises believe they’ve ‘approved’ AI vendors, but what they’ve actually approved is an interface, not the underlying system.” The credentials underneath the interface are the breach. ## Codex, where a branch name stole GitHub tokens [BeyondTrust](https://www.beyondtrust.com/blog/entry/openai-codex-command-injection-vulnerability-github-token) researcher Tyler Jespersen, with Fletcher Davis and Simon Stewart, found Codex cloned repositories using a GitHub OAuth token embedded in the git remote URL. During cloning, the branch name parameter flowed unsanitized into the setup script. A semicolon and a backtick subshell turned the branch name into an exfiltration payload. Stewart added the stealth. By appending 94 Ideographic Space characters (Unicode U+3000) after “main,” the malicious branch looked identical to the standard main branch in the Codex web portal. A developer sees “main.” The shell sees curl exfiltrating their token. OpenAI classified it Critical P1 and shipped full remediation by February 5, 2026. ## Claude Code, where two CVEs and a 50-subcommand bypass broke the sandbox [CVE-2026-25723](https://www.sentinelone.com/vulnerability-database/cve-2026-25723/) hit Claude Code’s file-write restrictions. Piped sed and echo commands escaped the project sandbox because command chaining was not validated. Patched in 2.0.55. [CVE-2026-33068](https://www.sentinelone.com/vulnerability-database/cve-2026-33068/) was subtler. Claude Code resolved permission modes from .claude/settings.json before showing the workspace trust dialog. A malicious repo set permissions.defaultMode to bypassPermissions. The trust prompt never appeared. Patched in 2.1.53. The 50-subcommand bypass landed last. Adversa found that Claude Code silently dropped deny-rule enforcement once a command exceeded 50 subcommands. Anthropic’s engineers had traded security for speed and stopped checking after the fiftieth. Patched in 2.1.90. “A significant vulnerability in enterprise AI is broken access control, where the flat authorization plane of an LLM fails to respect user permissions,” wrote Carter Rees, VP of AI and Machine Learning at [Reputation](https://reputation.com/) and a member of the Utah AI Commission. The repository decided what permissions the agent had. The token budget decided which deny rules survived. ## Copilot, where a pull request description and *[Content truncated...]* --- ## 2. Best AI Coding Agents in 2026, Ranked - MightyBot **URL:** [https://mightybot.ai/blog/coding-ai-agents-for-accelerating-engineering-workflows/](https://mightybot.ai/blog/coding-ai-agents-for-accelerating-engineering-workflows/) **Score:** 0.7785 ### Summary ## Key Trends Reshaping Coding AI in 2026 The terminal is the new battleground. Codex CLI, Claude Code, OpenCode, Gemini CLI, GitHub Copilot CLI, and Aider all compete in the terminal. The IDE is no longer the only surface for AI-assisted development — CLI agents offer deeper system access, scriptability, and integration with CI/CD pipelines. Multi-agent architectures are mainstream. Codex, Claude Code, Cursor, and Copilot all support spawning parallel sub-agents for complex tasks. A lead agent decomposes a problem, delegates subtasks, and merges results — enabling work that would overwhelm a single agent context. [...] MightyBot · Best AI Coding Agents in 2026, Ranked Summary: AI coding agents are autonomous developer tools that plan tasks, edit code across repositories, run tests, and submit pull requests with less human handholding than autocomplete assistants. In this April 2026 refresh, Codex moves to #1 because GPT-5.5 materially improves code quality and agentic execution inside OpenAI’s multi-surface coding workflow. Claude Code remains a top-tier terminal agent, but recent Claude Code harness issues make reliability and default reasoning controls part of the buying decision. Updated April 2026 ## The Top Coding AI Agents in 2026 ### Quick Comparison [...] FAQ ## Frequently Asked Questions What are the best AI coding agents in 2026? The top AI coding agents in 2026 are OpenAI Codex with GPT-5.5 for the best overall coding-agent workflow, Claude Code with Opus 4.7 for Claude-native terminal development, OpenCode for provider-agnostic open-source flexibility, Gemini CLI for free access to frontier models with 1M token context, and Cursor for the best AI-native IDE experience. How do AI coding agents improve software development workflows? ### Full Content [← Blog](/blog/) ai-agentsautomation # Best AI Coding Agents in 2026, Ranked The best AI coding agents of 2026, ranked by agentic coding performance, workflow depth, code quality, and real-world adoption. Codex, Claude Code, OpenCode, Gemini CLI, Cursor, and Copilot compared. MightyBot ·  **Summary:** AI coding agents are autonomous developer tools that plan tasks, edit code across repositories, run tests, and submit pull requests with less human handholding than autocomplete assistants. In this April 2026 refresh, Codex moves to #1 because GPT-5.5 materially improves code quality and agentic execution inside OpenAI’s multi-surface coding workflow. Claude Code remains a top-tier terminal agent, but recent Claude Code harness issues make reliability and default reasoning controls part of the buying decision. *Updated April 2026* ## The Top Coding AI Agents in 2026 ### **Quick Comparison** | Agent | Best For | Key Differentiator | | --- | --- | --- | | **Codex** | Overall best | GPT-5.5, 82.7% Terminal-Bench 2.0, multi-agent worktrees | | **Claude Code** | Claude-native terminal work | Opus 4.7, 1M context, strong SWE-Bench Pro, recent harness fixes | | **OpenCode** | Open source | 75+ LLM providers, fully offline | | **Gemini CLI** | Free access | 1M token context, free tier | | **Cursor** | AI-native IDE | Cloud agents, computer use | | **GitHub Copilot** | Ecosystem breadth | Async coding agent, widest IDE support | | **Devin** | Full autonomy | End-to-end sandboxed environment | | **Windsurf** | Large codebases | Cascade agent, Gartner Leader | | **Replit Agent** | Rapid prototyping | 200-min autonomous runs, free tier | ### **1. Codex (OpenAI) — Best Overall AI Coding Agent** [Codex](https://openai.com/codex/) is OpenAI’s coding agent platform spanning the Codex app, cloud delegation, an open-source CLI, IDE extensions, and connected ChatGPT workflows. With GPT-5.5 now available in Codex, it has become the strongest overall coding-agent choice for teams that want high code quality, multi-agent execution, background work, and human review in one system. **Key capabilities:** * **GPT-5.5 code quality** — OpenAI says GPT-5.5 is its strongest agentic coding model to date, with better system-level reasoning, debugging, testing, validation, and fewer tokens for the same Codex tasks * **Multi-agent worktrees** — Codex can run agents in parallel across cloud environments and local workspaces, letting teams split implementation, review, refactor, and test workstreams * **Local + cloud execution** — Developers can pair in the terminal or IDE, delegate longer work to the cloud, and continue from the same ChatGPT-connected account * **AGENTS.md convention** — Hierarchical project instructions guide codebase navigation, testing, review standards, and repo-specific engineering practices * **Automations and Skills** — Background scheduled agent runs and reusable skills make Codex useful beyond one-off code generation, including documentation, issue triage, CI/CD follow-up, and deployment workflows * **Code review quality** — Codex is increasingly differentiated by high-signal code review, test generation, and catching integration issues before humans merge changes **Why it’s #1:** Codex now has the best combination of model quality and product workflow. [GPT-5.5 reaches 82.7% on Terminal-Bench 2.0](https://openai.com/index/introducing-gpt-5-5/), outpacing Claude Opus 4.7 in OpenAI’s published comparison, and OpenAI says the gains are strongest in agentic coding, computer use, and long-running professional work. Claude still leads the public SWE-Bench Pro number, but Codex is the stronger overall system for teams managing real engineering work across CLI, IDE, cloud agents, worktrees, reviews, automations, and repo-specific standards. ### **2. Claude Code (Anthropic) — Best Claude-Native Terminal Coding Agent** [Claude Code](https://claude.com/product/claude-code) is Anthropic’s agentic coding tool that lives in your terminal, reads your codebase, edits files, runs commands, and manages git workflows through natural language. Powered by Claude Opus 4.7, it remains one of the strongest coding agents available, especially for developers who want a terminal-first Claude workflow and very large context. **Key capabilities:** * **Opus 4.7 for long-horizon coding** — [Anthropic says Opus 4.7 improves](https://www.anthropic.com/news/claude-opus-4-7) advanced software engineering, long-running tasks, instruction following, and self-verification over Opus 4.6 * **1M token context window** — Strong fit for monorepos, long sessions, and large codebase analysis * **Effort controls** — Claude Code exposes effort levels so developers can trade off reasoning depth, latency, and token cost; Opus 4.7 now defaults to xhigh for coding * **IDE + CLI + mobile workflows** — Works across terminal, editor integrations, and rem *[Content truncated...]* --- ## 3. CLAUDE.md, AGENTS.md, and Every AI Config File Explained - DEV Community **URL:** [https://dev.to/deployhq/claudemd-agentsmd-and-every-ai-config-file-explained-4pde](https://dev.to/deployhq/claudemd-agentsmd-and-every-ai-config-file-explained-4pde) **Score:** 0.7441 ### Summary DEV Community ## DEV Community DeployHQ Posted on Apr 27 • Originally published at deployhq.com on Mar 11 # CLAUDE.md, AGENTS.md, and Every AI Config File Explained Every AI coding tool now reads a configuration file from your project. Claude Code looks for `CLAUDE.md`. Codex CLI reads `AGENTS.md`. Gemini CLI checks for `GEMINI.md`. Cursor has `.cursorrules`. GitHub Copilot uses `copilot-instructions.md`. Windsurf has `.windsurfrules`. `CLAUDE.md` `AGENTS.md` `GEMINI.md` `.cursorrules` `copilot-instructions.md` `.windsurfrules` [...] | File | Tool | Location | Format | --- --- | | `CLAUDE.md` | Claude Code | Project root + `~/.claude/` | Markdown | | `AGENTS.md` | Codex CLI, Cursor, Claude Code (fallback) | Project root + subdirectories | Markdown | | `GEMINI.md` | Gemini CLI | Project root + `~/.gemini/` | Markdown | | `.cursorrules` | Cursor (legacy) | Project root | Plain text | | `.cursor/rules/.mdc` | Cursor (current) | `.cursor/rules/` directory | MDC (Markdown+) | | `.github/copilot-instructions.md` | GitHub Copilot | `.github/` directory | Markdown | | `.github/instructions/.instructions.md` | GitHub Copilot (scoped) | `.github/instructions/` | Markdown + frontmatter | | `.windsurfrules` | Windsurf (legacy) | Project root | Plain text | [...] `your-project/ ├── AGENTS.md ← Universal instructions (works with Codex, Cursor, Claude Code) ├── CLAUDE.md ← Claude-specific additions (if needed) ├── .github/ │ └── copilot-instructions.md ← Copilot-specific (if your team uses it) ├── .cursor/ │ └── rules/ ← Cursor-specific scoped rules (if needed) └── ...` Start with `AGENTS.md` as your single source of truth. It has the widest cross-tool support. Then add tool-specific files only when you need features that `AGENTS.md` can't provide (like Cursor's scoped activation or Copilot's glob patterns). `AGENTS.md` `AGENTS.md` For Claude Code users: you can make your `CLAUDE.md` simply reference the shared instructions: `CLAUDE.md` `Strictly follow the rules in ./AGENTS.md` ## Writing Effective Instructions ### Full Content  ## DEV Community        Posted on Apr 27 • Originally published at [deployhq.com](https://www.deployhq.com/blog/ai-coding-config-files-guide) on Mar 11 # CLAUDE.md, AGENTS.md, and Every AI Config File Explained Every AI coding tool now reads a configuration file from your project. [Claude Code](https://dev.to/deployhq/getting-started-with-claude-code-the-ai-coding-assistant-for-your-terminal-4cba) looks for `CLAUDE.md`. [Codex CLI](https://dev.to/deployhq/getting-started-with-openai-codex-cli-ai-powered-code-generation-from-your-terminal-5hm8) reads `AGENTS.md`. [Gemini CLI](https://dev.to/deployhq/getting-started-with-google-gemini-cli-open-source-ai-agent-for-your-terminal-25e1) checks for `GEMINI.md`. [Cursor](https://www.deployhq.com/guides/cursor) has `.cursorrules`. GitHub Copilot uses `copilot-instructions.md`. [Windsurf](https://www.deployhq.com/guides/windsurf) has `.windsurfrules`. `CLAUDE.md` `AGENTS.md` `GEMINI.md` `.cursorrules` `copilot-instructions.md` `.windsurfrules` These files solve a fundamental problem: AI models have no persistent memory. Every session starts blank. Without a configuration file, you'd repeat the same instructions every time — your tech stack, coding conventions, project structure, deployment rules. These files give your AI assistant the context it needs to produce useful code from the first prompt. This guide covers every major format, explains how they work, and shows you how to write effective instructions that actually improve your AI's output. ## The Complete Landscape Here's every AI configuration file format in use today: | File | Tool | Location | Format | | --- | --- | --- | --- | | `CLAUDE.md` | Claude Code | Project root + `~/.claude/` | Markdown | | `AGENTS.md` | Codex CLI, Cursor, Claude Code (fallback) | Project root + subdirectories | Markdown | | `GEMINI.md` | Gemini CLI | Project root + `~/.gemini/` | Markdown | | `.cursorrules` | Cursor (legacy) | Project root | Plain text | | `.cursor/rules/*.mdc` | Cursor (current) | `.cursor/rules/` directory | MDC (Markdown+) | | `.github/copilot-instructions.md` | GitHub Copilot | `.github/` directory | Markdown | | `.github/instructions/*.instructions.md` | GitHub Copilot (scoped) | `.github/instructions/` | Markdown + frontmatter | | `.windsurfrules` | Windsurf (legacy) | Project root | Plain text | | `.windsurf/rules/*.md` | Windsurf (current) | `.windsurf/rules/` directory | Markdown | `CLAUDE.md` `~/.claude/` `AGENTS.md` `GEMINI.md` `~/.gemini/` `.cursorrules` `.cursor/rules/*.mdc` `.cursor/rules/` `.github/copilot-instructions.md` `.github/` `.github/instructions/*.instructions.md` `.github/instructions/` `.windsurfrules` `.windsurf/rules/*.md` `.windsurf/rules/` Every tool has converged on the same core idea: a markdown file in your repository that the AI reads before doing anything. The differences are in naming and how they handle hierarchy. ## How Each Format Works ### CLAUDE.md — Claude Code Claude Code reads `CLAUDE.md` files from three locations, merged in order: `CLAUDE.md` `~/.claude/CLAUDE.md` `./CLAUDE.md` `CLAUDE.md` Claude Code also reads `AGENTS.md` as a fallback if no `CLAUDE.md` is found in a directory. This means if your team uses multiple AI tools, you can maintain a single `AGENTS.md` and it will work with Claude Code automatically. `AGENTS.md` `CLAUDE.md` `AGENTS.md` A key constraint: research suggests frontier LLMs can reliably follow around 150-200 instructions. Claude Code's system prompt already uses about 50 of those, so keep your `CLAUDE.md` concise — ideally under 300 lines. `CLAUDE.md` ### AGENTS.md — The Cross-Tool Standard `AGENTS.md` is the closest thing to a universal standard. Originally popularised by OpenAI's Codex CLI, it's now an [open format stewarded by the Linux Foundation](https://agents.md/) with adoption across 60,000+ open-source projects. It's supported by: `AGENTS.md` `.cursorr *[Content truncated...]* --- ## 4. Claude Opus 4.7, GPT-5.5, Gemini-3.1, Cursor AI, Copilot, Codex, Cline, and ChatGPT, AI Copilot, AI Agents and Debugger, Code Assistants, Code Chat, Code Generator, Generative AI, Code Completion,Aut - Visual Studio Marketplace **URL:** [https://marketplace.visualstudio.com/items?itemName=Sixth.sixth-ai](https://marketplace.visualstudio.com/items?itemName=Sixth.sixth-ai) **Score:** 0.6887 ### Summary | Sixth AIThe AI Coding Agent That Works While You Sleep Install from VS Marketplace | Discord | Twitter | Website | VS Marketplace Downloads Sixth AI is an autonomous coding agent that lives inside VS Code. Give it a task, and it will plan, write, edit, and debug code across your entire project — then let you review every change before it's applied. No copy-pasting. No context-switching. Just describe what you want and watch it build. Unlike basic copilots that suggest one line at a time, Sixth AI thinks through problems, spins up sub-agents for parallel work, and can even be controlled remotely from your phone via Telegram. It's the closest thing to having a senior engineer pair-programming with you 24/7. --- Edit Files Across Your Entire Project Sixth AI doesn't just suggest [...] # Sixth AI ### The AI Coding Agent That Works While You Sleep Install from VS Marketplace | Discord | Twitter | Website | VS Marketplace Downloads VS Marketplace Downloads Sixth AI is an autonomous coding agent that lives inside VS Code. Give it a task, and it will plan, write, edit, and debug code across your entire project — then let you review every change before it's applied. No copy-pasting. No context-switching. Just describe what you want and watch it build. Unlike basic copilots that suggest one line at a time, Sixth AI thinks through problems, spins up sub-agents for parallel work, and can even be controlled remotely from your phone via Telegram. It's the closest thing to having a senior engineer pair-programming with you 24/7. ### Edit Files Across Your Entire Project [...] command Works with Any Model — use Claude, GPT, Gemini, DeepSeek, or bring your own API key Supported Languages & Frameworks JavaScript, TypeScript, Python, Java, C, C++, C#, Go, PHP, Ruby, Kotlin, Dart, Rust, Swift, SQL, Shell, R, Julia, Lua, Perl, React, Vue, Svelte, HTML, CSS, Tailwind, Terraform, YAML, Docker, and more. --- Get Started in 30 Seconds 1. Install Sixth AI from the VS Marketplace 2. Open the Sixth AI panel in your sidebar 3. Describe what you want to build, fix, or change 4. Review the agent's plan and approve the edits 5. Ship it --- Community & Support Discord — get help, share feedback, and connect with other developers Twitter — follow for updates and tips Website — learn more about Sixth AI --- License Apache 2.0 © 2026 SixHq Inc. | | ### Full Content  | | | | --- | --- | | Claude Opus 4.7, GPT-5.5, Gemini-3.1, Cursor AI, Copilot, Codex, Cline, and ChatGPT, AI Copilot, AI Agents and Debugger, Code Assistants, Code Chat, Code Generator, Generative AI, Code Completion,Aut | Claude Opus 4.7, GPT-5.5, Gemini-3.1, Cursor AI, Copilot, Codex, Cline, and ChatGPT, AI Copilot, AI Agents and Debugger, Code Assistants, Code Chat, Code Generator, Generative AI, Code Completion,Aut [Sixth](publishers/Sixth) [trysixth.com](https://trysixth.com) | 346,932 installs [| (44)](#review-details) | Free Claude Opus 4.7, GPT-5.5, Gemini-3.1, AI Coding Assistant is a lightweight for helping developers automate all the boring stuff like writing code, real-time code completion, debugging, auto generating doc string and many more. Trusted by 100K+ devs from Amazon, Apple, Google, & more. Offers all the [Install](vscode:extension/Sixth.sixth-ai) [Trouble Installing?](https://aka.ms/vscode_extn_install) |  # Claude Opus 4.7, GPT-5.5, Gemini-3.1, Cursor AI, Copilot, Codex, Cline, and ChatGPT, AI Copilot, AI Agents and Debugger, Code Assistants, Code Chat, Code Generator, Generative AI, Code Completion,Aut ## [Sixth](publishers/Sixth)      | | | | | | | | | | | | | | | | | | | | | | | | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | Sixth AIThe AI Coding Agent That Works While You Sleep [**Install from VS Marketplace**](https://marketplace.visualstudio.com/items?itemName=Sixth.sixth) | [**Discord**](https://discord.gg/GKEasakXKc) | [**Twitter**](https://twitter.com/sixth_hq) | [**Website**](https://trysixth.com) | VS Marketplace Downloads Sixth AI is an autonomous coding agent that lives inside VS Code. Give it a task, and it will plan, write, edit, and debug code across your entire project — then let you review every change before it's applied. No copy-pasting. No context-switching. Just describe what you want and watch it build. Unlike basic copilots that suggest one line at a time, Sixth AI **thinks through problems**, spins up sub-agents for parallel work, and can even be controlled remotely from your phone via Telegram. It's the closest thing to having a senior engineer pair-programming with you 24/7. --- Edit Files Across Your Entire Project Sixth AI doesn't just suggest code — it **writes and edits files directly** in your workspace. Tell it to add a feature, fix a bug, or refactor a module, and it will create, modify, and delete files as needed, presenting you with a clean diff to approve. * **Multi-file edits** in a single conversation — no more juggling tabs * **Automatic error detection** — catches missing imports, syntax errors, and type mismatches as it works * **Full diff review** — see exactly what changed before you accept anything * **Timeline tracking** — every change is recorded, so you can always roll back --- Deep Planning Mode Complex tasks need a plan before code. Sixth AI's **Deep Planning Mode** breaks down your request into a structured, step-by-step strategy before writing a single line — so the agent doesn't just code fast, it codes *right*. * **Architectural thinking** — analyzes your codebase structure, dependencies, and patterns before proposing changes * **Step-by-step roadmap** — see the full plan and adjust it before execution begins * **Fewer dead ends** — planning up front means less wasted time on approaches that don't work * **Perfect for big tasks** — feature builds, refactors, and migrations that span dozens of files --- Sub-Agents for Parallel Work Why do one thing at a time when you can do five? Sixth AI can **spin up sub-agents** that work on different parts of your task simultaneously — researching, writing, and validating in parallel. * **Parallel execution** — multiple agents tackle different files or concerns at once * **Divide and conquer** — the main agent orchestrates while sub-agents handle the details * **Faster results** — tasks that used to take 10 minutes now finish in 2 * **Smart coordination** — sub-agents share context so their work doesn't conflict --- Remote Control via Telegram Start a task on yo *[Content truncated...]* --- ## 5. GitHub - yzhao062/agent-style: 21 writing rules for AI coding and ... **URL:** [https://github.com/yzhao062/agent-style](https://github.com/yzhao062/agent-style) **Score:** 0.6804 ### Summary | Tool | install\_mode | Target path | --- | Claude Code | `import-marker` | `CLAUDE.md` (marker block with `@.agent-style/claude-code.md`) | | AGENTS.md (Codex, Jules, Zed, Warp, Gemini CLI, VS Code, and others) | `append-block` | `AGENTS.md` at repo root | | GitHub Copilot (repo-wide) | `append-block` | `.github/copilot-instructions.md` | | GitHub Copilot (path-scoped) | `owned-file` | `.github/instructions/agent-style.instructions.md` | | Cursor | `owned-file` | `.cursor/rules/agent-style.mdc` | | Anthropic Skills | `owned-file` | `.claude/skills/agent-style/SKILL.md` | | Codex (API / manual paste) | `print-only` | `.agent-style/codex-system-prompt.md` (user pastes into API system prompt) | [...] `CHANGELOG.md` ↑ back to top ## About 21 writing rules for AI coding and writing agents. Drop-in for Claude Code, Codex, Copilot, Cursor, and Aider, so their output reads like a tech pro. ### Topics ### Resources ### Uh oh! There was an error while loading. Please reload this page. There was an error while loading. Please reload this page. ### Stars ### Watchers ### Forks ## Releases 7 ## Packages 0 ### Uh oh! There was an error while loading. Please reload this page. There was an error while loading. Please reload this page. ## Contributors ### Uh oh! There was an error while loading. Please reload this page. There was an error while loading. Please reload this page. ## Languages ## Footer ### Footer navigation [...] Install the CLI once; then pick the paths you want. ### 1. Soft Enforcement — rules at generation time Once the CLI is installed (per the install block above), paste the block below into your AI agent — it picks the right adapter, runs `enable`, verifies, and offers `style-review` if your host supports skills. Works with any AI agent that can run shell commands: Claude Code, Codex, Cursor, Aider, Kiro, GitHub Copilot, etc. ### Full Content ## Navigation Menu # Search code, repositories, users, issues, pull requests... # Provide feedback We read every piece of feedback, and take your input very seriously. # Saved searches ## Use saved searches to filter your results more quickly To see all available qualifiers, see our [documentation](https://docs.github.com/search-github/github-code-search/understanding-github-code-search-syntax). # yzhao062/agent-style ## Folders and files | Name | | Name | Last commit message | Last commit date | | --- | --- | --- | --- | --- | | Latest commit History[53 Commits](/yzhao062/agent-style/commits/main/) 53 Commits | | | | [.github](/yzhao062/agent-style/tree/main/.github ".github") | | [.github](/yzhao062/agent-style/tree/main/.github ".github") | | | | [LICENSES](/yzhao062/agent-style/tree/main/LICENSES "LICENSES") | | [LICENSES](/yzhao062/agent-style/tree/main/LICENSES "LICENSES") | | | | [agents](/yzhao062/agent-style/tree/main/agents "agents") | | [agents](/yzhao062/agent-style/tree/main/agents "agents") | | | | [docs](/yzhao062/agent-style/tree/main/docs "docs") | | [docs](/yzhao062/agent-style/tree/main/docs "docs") | | | | [enforcement](/yzhao062/agent-style/tree/main/enforcement "enforcement") | | [enforcement](/yzhao062/agent-style/tree/main/enforcement "enforcement") | | | | [packages](/yzhao062/agent-style/tree/main/packages "packages") | | [packages](/yzhao062/agent-style/tree/main/packages "packages") | | | | [scripts](/yzhao062/agent-style/tree/main/scripts "scripts") | | [scripts](/yzhao062/agent-style/tree/main/scripts "scripts") | | | | [skills/style-review](/yzhao062/agent-style/tree/main/skills/style-review "This path skips through empty directories") | | [skills/style-review](/yzhao062/agent-style/tree/main/skills/style-review "This path skips through empty directories") | | | | [.gitignore](/yzhao062/agent-style/blob/main/.gitignore ".gitignore") | | [.gitignore](/yzhao062/agent-style/blob/main/.gitignore ".gitignore") | | | | [.markdownlint-cli2.jsonc](/yzhao062/agent-style/blob/main/.markdownlint-cli2.jsonc ".markdownlint-cli2.jsonc") | | [.markdownlint-cli2.jsonc](/yzhao062/agent-style/blob/main/.markdownlint-cli2.jsonc ".markdownlint-cli2.jsonc") | | | | [CHANGELOG.md](/yzhao062/agent-style/blob/main/CHANGELOG.md "CHANGELOG.md") | | [CHANGELOG.md](/yzhao062/agent-style/blob/main/CHANGELOG.md "CHANGELOG.md") | | | | [NOTICE.md](/yzhao062/agent-style/blob/main/NOTICE.md "NOTICE.md") | | [NOTICE.md](/yzhao062/agent-style/blob/main/NOTICE.md "NOTICE.md") | | | | [README.md](/yzhao062/agent-style/blob/main/README.md "README.md") | | [README.md](/yzhao062/agent-style/blob/main/README.md "README.md") | | | | [RELEASING.md](/yzhao062/agent-style/blob/main/RELEASING.md "RELEASING.md") | | [RELEASING.md](/yzhao062/agent-style/blob/main/RELEASING.md "RELEASING.md") | | | | [RULES.md](/yzhao062/agent-style/blob/main/RULES.md "RULES.md") | | [RULES.md](/yzhao062/agent-style/blob/main/RULES.md "RULES.md") | | | | [SOURCES.md](/yzhao062/agent-style/blob/main/SOURCES.md "SOURCES.md") | | [SOURCES.md](/yzhao062/agent-style/blob/main/SOURCES.md "SOURCES.md") | | | | [TODO.md](/yzhao062/agent-style/blob/main/TODO.md "TODO.md") | | [TODO.md](/yzhao062/agent-style/blob/main/TODO.md "TODO.md") | | | | [adapter-matrix.md](/yzhao062/agent-style/blob/main/adapter-matrix.md "adapter-matrix.md") | | [adapter-matrix.md](/yzhao062/agent-style/blob/main/adapter-matrix.md "adapter-matrix.md") | | | | View all files | | | ## Latest commit ## History ## Repository files navigation # The Elements of Agent Style **Make your AI agent write like a tech pro.** *21 rules: 12 from classic writing guides, 9 from field observation of LLM output, 2022-2026.* [](https://pypi.org/project/agent-style/) [](https://www.npmjs.com/package/agent-style) [](/yzhao062/agent-style/blob/main/NOTICE.md) [](https://github.com/yzhao062/agent-style/actions/workflows/validate.yml) [![GitHub stars](https://camo.githubusercontent.com/d9d50f3c15d5c3bce4a2a5da6ca446ee678350ade088d7c8a3e054c47a75782f/68747470733a2f2f696d672e736869656c64732e696f2f6769746875622f73746 *[Content truncated...]* ---