AIQB
TutorialsOrdinary

One Markdown File Can Hijack Your AI Coding Agent: The Hidden Cost of Blindly Trusting AGENTS.md

Source: DEV Community·
Claude

Summary

How a single AGENTS.md or CLAUDE.md file in your Downloads folder or a pull request can silently hijack your AI coding agent, steal OAuth tokens, and execute shell commands—plus how to lock down your local workflow.
CategoryAI Tutorials & Practice
TierOrdinary
Published
Indexed by AIQB
SourceDEV Community
AIQB record IDintel-6fe366c95ea5e5fd43b84696
Read original source ↗More Tutorials

Related stories

Matched by category, entity tags and source

TutorialsOrdinary
We instrumented Claude Code for our team. Here's what the data showed
AI coding only started paying off for my team when we stopped treating it as a tool everyone uses...
Source: DEV Community · Claude
TutorialsOrdinary
Claude Code wrote almost all of my app. My job turned into saying no.
Three weeks ago I started a desktop app. Since then: 169 commits, about 48,000 lines in src/, and I...
Source: DEV Community · Claude
TutorialsOrdinary
Your Claude Code allowlist is wider than you think: building an auditor for the rules you forgot you wrote
Every Claude Code user has the same file, and nobody reads it. settings.json accumulates permission...
Source: DEV Community · Claude
TutorialsOrdinary
What I learned building a lossless token compressor for Claude Code
In a long Claude Code session, most of what gets sent to the API isn't your prompts. It's tool...
Source: DEV Community · Claude
TutorialsOrdinary
49 minutes from first commit to production: building a planning poker app with Claude Code
My team estimates stories with planning poker. For a long time we used a free site that was covered...
Source: DEV Community · Claude
TutorialsOrdinary
I built a Claude Code skill that makes video ads without a video model
How I built reel-studio, a Claude Code skill that makes Reels and TikTok ads: deterministic HTML motion, Playwright frames, ffmpeg loudness, Whisper captions.
Source: DEV Community · Claude
TutorialsOrdinary
Safe Tools, Unsafe Session: Why Agent Permissions Need History
Taint tracking for AI agent tool calls: per-tool allow/deny rules miss the lethal trifecta, so TaintGuard gates Claude Code by what the session has read.
Source: DEV Community · Claude
TutorialsOrdinary
How KV Cache Makes LLMs Fast, and Why It Eats GPU Memory
Every LLM you've used, ChatGPT, Claude, Gemini, Llama, does the same thing under the hood. It...
Source: DEV Community · ChatGPT / Claude / Gemini
TutorialsOrdinary
Crysto: how I run Claude, Codex and other AI models as one team. The models were the easy part.
What went wrong when my AI agents started sharing projects, and the boring text files that fixed it.
Source: DEV Community · Claude
TutorialsOrdinary
Claude Reads Screenshots at 2,576px Now. The Same PNG Can Cost 3x the Tokens
Claude 4.7 and later read images up to 2,576px, so a big screenshot can cost 3x the tokens. The 28px patch math, plus a TypeScript script to count yours.
Source: DEV Community · Claude