Unverified50% confidenceEventExact time
Forever Security reported on October 22 that in a real customer's Downloads folder it found an agents.md file disguised as engineering notes, which instructed the coding agent to add a server to its own configuration and install a malicious package named FunctionColor from NPM, and that its product blocked this behavior before code execution.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/31/2026
First Seen
Valid until: 11/14/2026
Sources
AI Weekly: GPT-6.1 Terra Enters at Low Price, AGENTS.md Becomes a New Attack Surface
youtubeAI Without the Noise10/29/2026
Related Entities
Related Claims
UnverifiedIn August, a virus-infected version of the NX tool was published to NPM, and the malicious code sought out AI agents such as Claude Code and Gemini CLI, asked them to scan for passwords and keys, and enabled Dangerously Skip Permissions; this ultimately led to the leakage of 2,349 keys and tokens, with AI executing the task on 95 machines.63% similarUnverifiedAttackers can hide malicious code in postinstall scripts, causing it to execute in the user's environment when downstream projects install dependencies, while remaining transparent to approval processes that only check package names and version numbers63% similarUnverifiedPrompt injection is a security vulnerability unique to LLM applications, where attackers can embed malicious instructions in tool results returned by an MCP server to manipulate agent behavior.62% similarUnverifiedThe commit-rewriter project originated from the author's need to clean up redundant content generated by programming agents and private repository issue IDs in early commits while preparing a Datasette security update for public release.62% similarUnverifiedInvariant Labs demonstrated a tool poisoning attack: hidden instructions in the tool description of a "daily fun fact" MCP server manipulated the model into extracting WhatsApp history and sending it to the attacker.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1119330API
curl https://kongchang.com/api/v1/knowledge/claims/1119330MCP
get_claim(id=1119330)