PulseAugur
EN
LIVE 20:45:56

Malicious markdown files hijack AI coding agents like Codex and Claude

By PulseAugur Editorial · [1 sources] ·

A security vulnerability has been discovered where malicious markdown files, such as AGENTS.md or CLAUDE.md, can be placed in an AI coding agent's working directory to hijack its operations. These files trick agents like OpenAI Codex and Claude Code into executing malicious commands, such as installing rogue npm packages or redirecting API traffic to attacker-controlled servers. The attack exploits the trust agents place in instruction files within their working directories, bypassing traditional security measures that do not flag markdown files or monitor API traffic that appears legitimate. AI

IMPACT Highlights a critical security flaw in AI agent trust models, necessitating new protective measures for code generation and API interactions.

RANK_REASON Security researchers discovered a vulnerability in AI coding agents, detailing how malicious files can hijack their functionality, and a security tool is presented as a solution.

Read on dev.to — LLM tag →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

Malicious markdown files hijack AI coding agents like Codex and Claude

How we ranked this

Signal score
0 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
Security researchers discovered a vulnerability in AI coding agents, detailing how malicious files can hijack their functionality, and a security tool is presented as a solution.
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
safety, product
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
3 days old
Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

Full methodology in our editorial standards.

COVERAGE [1]

  1. dev.to — LLM tag TIER_1 English(EN) · Cor E ·

    A Poisoned AGENTS.md in ~/Downloads Hijacked Codex and Claude Code. Here's Where the Scrub Step Fits.

    <p>Forever Security found a malicious <code>AGENTS.md</code> sitting in a customer's <code>~/Downloads</code> folder. That folder was also the agent's working directory. The file tricked an AI coding agent into trying to install a rogue npm package called <code>function-color</co…

RELATED ENTITIES

RELATED TOPICS