Malicious code in @ch4acko3/frontal-lobe (npm)
The npm package @ch4acko3/frontal-lobe contained malicious code that exfiltrated AI session data, including user prompts, model outputs, and source-code context, to a hardcoded IP endpoint via plaintext HTTP. The postinstall script automatically enabled data collection without explicit user consent.
- Disclosed
- Last updated
- Blast radius
- All users who installed @ch4acko3/frontal-lobe package
- Ecosystems
- Attack vectors
- Affected entities
- @ch4acko3/frontal-lobenpm package containing malicious code
The npm package @ch4acko3/frontal-lobe was found to contain malicious code that exfiltrated sensitive AI session data. The package hardcoded an endpoint (http://47.112.15.137:8789/frontal/trace-sync) and uploaded full Codex/Claude AI session rollouts, including user prompts, model outputs, and source-code context, as gzip+base64 encoded chunks.\n\nThe postinstall lifecycle script (bin/frontal-codex-plugin-postinstall.mjs) automatically created a configuration file at ~/.frontal_lobe/config.toml with trace synchronization enabled by default, set to run every 30 minutes. The script prompted users for a Memorax user_id and silently reused it as the trace-sync identity. A background gateway process was started to periodically flush collected data.\n\nThe exfiltration endpoint was a bare IPv4 address over plaintext HTTP, not the documented test service (test-code.beta.memorax.net), and was activated at install time without a distinct opt-in mechanism. The seeded configuration's internal label describing this as an approved test did not constitute user consent.\n\nThe incident was identified by Amazon Inspector and credited to the OpenSSF malicious-packages project.
Indicators of compromise
- Packages
- @ch4acko3/frontal-lobe
- IPs
- 47.112.15.137
Remediation
- Immediately uninstall @ch4acko3/frontal-lobe from all systems
- Audit ~/.frontal_lobe/config.toml and related directories for evidence of data exfiltration
- Review network logs for connections to 47.112.15.137:8789 to determine exposure scope
- Assume any AI session data (prompts, outputs, source code) processed while the package was installed may have been exfiltrated
- Rotate any credentials or sensitive information that may have been included in AI prompts or context
- Do not reinstall this package; use alternative, verified packages for required functionality
Sources
- GitHub Advisory GHSA-m8jr-68xh-8qp9 · GitHub Advisory Database
Cite this entry
"Malicious code in @ch4acko3/frontal-lobe (npm)." supplychainattack.org, Supply Chain Attack Incident Catalog. Disclosed August 7, 2026; last updated August 7, 2026. https://supplychainattack.org/incident/malicious-code-in-ch4acko3-frontal-lobe-npm-1ro6iz
Suggest a correction
Found an error or have a newer source? Corrections to factual errors take priority over new entries.
Related incidents
- containedcritical
Malicious code in 3-buildsight-web (npm)
The npm package 3-buildsight-web contains malicious code that fetches and executes arbitrary JavaScript from a remote Bitbucket URL on module load, granting full page privileges to an unverified third party.
npmCompromised package - resolvedcritical
Malicious code in @usaa-grp-personal-profile/personal-profile-common (npm)
Malicious package @usaa-grp-personal-profile/personal-profile-common version 999.0.0 published to public npm as a dependency-confusion attack targeting an organization's internal registry. The package contains preinstall and postinstall hooks that exfiltrate hostname, username, and working directory to an attacker-controlled OAST endpoint on every install.
npmDependency confusionCompromised package - resolvedcritical
Malicious code in create-coin (npm)
create-coin@20.1.1 on npm contained malicious postinstall code that collected sensitive environment variables, credentials, and host reconnaissance data. The package impersonated an esbuild platform build and exfiltrated data to a hardcoded endpoint.
npmCompromised packageTyposquatting - resolvedcritical
Malicious code in @syncraft-labs/vue (npm)
The npm package @syncraft-labs/vue version 0.4.1 contained malicious obfuscated code in its ESM build that implements a blockchain-hosted C2 dropper (EtherHiding pattern), enabling remote code execution on any consumer importing the package.
npmCompromised package