registry / @brimveyn/aimux / 1.18.1

@brimveyn/aimux@1.18.1

A terminal multiplexer for AI CLIs. Run Claude, Codex, OpenCode side-by-side with tabbed navigation, split panes, and persistent sessions.

AI Security Review

scanned 12h ago · by lpm-firewall-ai

LPM treats this as warn-only first-party agent extension lifecycle risk. No install-time execution or confirmed exfiltration chain was found. At runtime, opt-in features can modify Claude Code hook settings and access local AI CLI credentials to request usage data from provider-aligned APIs.

Static reason

No blocking static signals were detected.

Trigger

User runs Aimux with `integrations.claudeHooks` or `statusBar.aiUsage.enabled` configured.

Impact

Can alter user Claude hook configuration and transmit provider credentials only to the corresponding official usage endpoints.

Mechanism

Opt-in Claude hook integration and authenticated AI usage polling.

Rationale

The package has a real but opt-in AI-agent extension and credential-handling surface, warranting a warning under the firewall policy. Source inspection found no concrete malicious chain, lifecycle abuse, remote payload execution, or unconsented install-time control-surface mutation.

Evidence

package.jsonsrc/app.tsxsrc/config/loader.tssrc/integrations/claude-hooks-install.tssrc/services/ai-usage/adapters/claude.tssrc/services/ai-usage/adapters/codex.ts~/.claude/settings.json~/.codex/auth.json~/.local/state/aimux-<profile>/claude-hook.url

Network endpoints4

api.anthropic.com/api/oauth/usage

@brimveyn/aimux@1.18.1

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code