Case agents-md-hijack-rogue-npm-packageAgent hijackWorldwide
Malicious AGENTS.md file in Downloads folder makes AI coding agent try to install rogue npm package
Forever Security researchers found a malicious AGENTS.md file in a customer's ~/Downloads folder, which was also the agent's working directory, that led an AI coding agent to attempt installing a malicious npm package called 'function-color'. The researchers also built a proof of concept showing a web page could silently drop such a file into Downloads.
MediumAttack on AIStatus: Under investigationVerified case file
Open in the interactive mapCite this case
- Reported
- Oct 25, 2026
- Location
- Worldwide
- Target
- AI coding agent on a customer's machine
- Severity
- Medium
- Verification
- Reported by credible press or researchers
- Triggered by
- Unknown
- Actor type
- Unknown
- Sector
- Unknown
- Reported loss
- None reported
- CVE
- None
What happened
Forever Security researchers found a malicious AGENTS.md file in a customer's ~/Downloads folder, which was also the agent's working directory, that led an AI coding agent to attempt installing a malicious npm package called 'function-color'. The researchers also built a proof of concept showing a web page could silently drop such a file into Downloads.
Impact: The agent installed a rogue npm package on a customer's system. The extent of compromise and any data loss are not stated.
Who triggered it
Actor: Unknown
Attribution: Unknown. The threat actor is not identified in the available text; the finding was reported by researchers via BleepingComputer.
Motive: Unknown
How it happened
Attack vector: Malicious AGENTS.md instruction file placed in the Downloads folder, read by the AI coding agent as trusted instructions.
Root cause: The coding agent treated an untrusted AGENTS.md file as legitimate instructions, which allowed indirect prompt injection and installation of a package.
Timeline
- Reported by BleepingComputer
Response
Researchers identified and reported the file. No vendor or customer remediation is described in the summary.
Current status: Under investigation
Research warnings this case matches
- Seen in real incidents AI systems themselves become targets · International AI Safety Report
- Seen in real incidents Privacy and security · MIT AI Risk Repository
Compared with published AI risk research from several institutes on the Research vs reality page.
Sources (open in a new tab)
- BleepingComputer bleepingcomputer.com
- Help Net Security helpnetsecurity.com
Cite this case
Malicious AGENTS.md file in Downloads folder makes AI coding agent try to install rogue npm package. AI Threat Watch, case agents-md-hijack-rogue-npm-package, Oct 25, 2026. https://aithreat.io/cases/agents-md-hijack-rogue-npm-package/
Permanent link: https://aithreat.io/cases/agents-md-hijack-rogue-npm-package/
Related cases
The same type of incident, most recent first.