triflux@10.80.1
CLI-first multi-model orchestrator for Claude Code — route tasks to Codex, Antigravity, and Claude
AI Security Review
scanned 5d ago · by lpm-firewall-aiThe postinstall hook performs an unprompted change to global Claude agent settings, creating residual agent setup risk. The inspected source does not establish a concrete attack that warrants blocking publication.
Decision evidence
public snapshot- package.json runs scripts/setup.mjs from postinstall.
- The postinstall branch runs setup for an installed package; the checkout guard only skips non-node_modules checkouts.
- Setup changes the user’s Claude settings to enable experimental agent teams and automatic teammate mode, then persists those changes before consent steps.
- Consent steps separately handle optional installs and Codex hook approval; they do not gate those settings changes.
- The inspected TUI source formats findings and sanitizes text; these runtime-operation hints show no attack behavior.
- The postinstall guard avoids changing the real home for checkout installs.
- The inspected files show no network recipient or credential transfer associated with the settings change.
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@10.80.1.
See all versions of triflux for other recorded verdicts.
- Review the evidence and your use of triflux@10.80.1 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.