@ai-support-agent/cli
AI Support Agent CLI client
This report shows the evidence for @ai-support-agent/cli@0.4.2. 29 of 46 known versions are blocked or flagged. The newest version, 0.5.16, is allowed. See all versions.
OpenSSF/OSV Malicious Advisory
imported 3mo ago · by OpenSSF/OSVOpenSSF/OSV advisory MAL-2026-10707 confirms this npm version as malicious.
Advisory
MAL-2026-10707
Source
OpenSSF Malicious Packages via OSV
Summary
Malicious code in @ai-support-agent/cli (npm)
Details
When `ai-support-agent start` is invoked, the CLI subscribes to a remote AppSync/WebSocket channel at api.ai-support-agent.com and dispatches server-delivered messages into local execution sinks. `execute_command` messages are passed to `spawn(shellCmd, ['-c', command])` in shell-executor.js; `terminal-ws` `stdin` frames are base64-decoded and written into a node-pty session (terminal-websocket.js `handleStdin`: `Buffer.from(msg.data,'base64').toString('utf-8'); session.write(decoded)`); additional handlers cover file_read/write/delete, process_kill, ssh_exec, server_setup_exec, reboot, and update. Whoever controls the vendor server — or anyone who obtains a valid agent token — has full shell-level control of the host running the agent. Related components include dist/vscode/vscode-server.js and dist/browser/browser-local-server.js (local HTTP endpoints and ping-based reachability probes) and dist/cli/service/{darwin,linux}-service.js (installs the agent as a persistent system service that modifies PATH), which extend and persist that control surface.
References
Independent scan evidence
public snapshotBehavioral surface
ChildProcessCryptoDynamicRequireEnvironmentVarsFilesystemNetworkWebSocket
HighEntropyStringsUrlStrings
Source & flagged code
8 flaggedansible/roles/database/tasks/main.ymlView file
85patternName = generic_password
severity = medium
line = 85
matchedText = password... }}"
Medium
Secret Pattern
Package contains a possible secret pattern.
ansible/roles/database/tasks/main.ymlView on unpkg · L8589patternName = generic_password
severity = medium
line = 89
matchedText = login_pa... }}"
Medium
Secret Pattern
Hardcoded password in ansible/roles/database/tasks/main.yml
ansible/roles/database/tasks/main.ymlView on unpkg · L89119patternName = generic_password
severity = medium
line = 119
matchedText = password... }}"
Medium
Secret Pattern
Hardcoded password in ansible/roles/database/tasks/main.yml
ansible/roles/database/tasks/main.ymlView on unpkg · L119Findings
2 High9 Medium5 Low
HighTrigger Reachable Persistencedist/terminal/terminal-session.js
HighPrevious Version Dangerous Deltadist/browser/playwright-subprocess-executor.js
MediumSecret Patternansible/roles/database/tasks/main.yml
MediumDynamic Requiredist/ws-reconnect.js
MediumNetwork
MediumEnvironment Vars
MediumInstall Persistencedist/terminal/terminal-session.js
MediumShips Build Helperdocker/bashrc-extra.sh
MediumStructural Risk Force Deep Review
MediumSecret Patternansible/roles/database/tasks/main.yml
MediumSecret Patternansible/roles/database/tasks/main.yml
LowNon Install Lifecycle Scripts
Low
Affected versions and remediation
This report applies to @ai-support-agent/cli@0.4.2.
The versions list below shows the verdict for each recorded version.
- Avoid installing @ai-support-agent/cli@0.4.2. Remove it from direct dependencies and check your lockfile for transitive copies.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
- If this version ran, investigate the affected machine and build environment. Rotate credentials it could access and rebuild from a trusted environment.
Evidence last updated: .
Source advisory published: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.