triflux@10.74.0
CLI-first multi-model orchestrator for Claude Code — route tasks to Codex, Antigravity, and Claude
AI Security Review
scanned 9d ago · by lpm-firewall-aiThe postinstall performs unprompted first-party agent setup by syncing files and enabling Claude Agent Teams. Codex hooks remain untrusted until approval, and the inspected source does not establish a separate malicious chain, so this warrants a warning rather than a block.
Decision evidence
public snapshot- package.json runs scripts/setup.mjs from postinstall, and the installed script enters runDeferred.
- Setup syncs package-owned files into the user's .claude directory.
- The setup path enables Claude Agent Teams and teammateMode in the user's settings.
- Codex hooks are initially configured without trust; noninteractive setup defers approval, and trust requires an explicit approval path.
- Other inspected flagged locations identify configuration or local diagnostics paths; those excerpts establish no separate attack behavior.
- The setup code keeps Codex hooks untrusted unless approval is given; the inspected source does not establish credential theft or a separate malicious chain.
Source & flagged code
25 flagged · loading sourcePackage defines install-time lifecycle scripts.
package.jsonView on unpkgPackage source references shell execution.
Findings
Affected versions and remediation
This report applies to triflux@10.74.0.
See all versions of triflux for other recorded verdicts.
- Review the evidence and your use of triflux@10.74.0 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.