@livx.cc/agentx@0.99.111
Edge-native AI agent runtime — drives a virtual filesystem via any LLM (ai.libx.js). Same bytes run in node, browser, or edge.
AI Security Review
scanned 4d ago · by lpm-firewall-aiThe source confirms conditional credential forwarding to an unrestricted configured destination, which warrants a warning. The default destination is Bodify and the request is gated on an available token, so the inspected behavior does not support a publish block.
Decision evidence
public snapshot- The manifest’s prepare script sets a Git hooks path, and prepublishOnly runs the build.
- The CLI loads environment files from `.agent`, gets a Bodify key from the environment or Para Keychain, and sends it as a bearer credential to `BODIFY_URL`; the default destination is Bodify.
- The Bodify request runs during CLI startup when a key is available, and the configured destination has no host allowlist.
- The cited `.claude/worktrees` code supports worktree isolation; runtime instruction loading reads project instruction files.
- The CLI’s `/init` command scaffolds `AGENTS.md` only when invoked and leaves existing `AGENTS.md` or `CLAUDE.md` files untouched.
- The shell source describes an OS sandbox with constrained writes, and the MCP client starts the configured server command.
- The credential request is for Bodify secrets retrieval and defaults to `https://bodify.bod.ee`; exposure depends on configuring `BODIFY_URL` to another destination.
- The inspected agent integration locations do not establish an automatic broad agent configuration change or credential transfer.
Source & flagged code
12 flagged · loading sourcePackage source references child process execution.
dist/tools.shell.jsView on unpkg · L891Findings
Affected versions and remediation
This report applies to @livx.cc/agentx@0.99.111.
See all versions of @livx.cc/agentx for other recorded verdicts.
- Review the evidence and your use of @livx.cc/agentx@0.99.111 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.