@livx.cc/agentx@0.99.101
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 forwarding of a saved service token to an unrestricted configurable destination, so this warrants a warning for credential exposure. The default destination and stated secrets lookup are Bodify-aligned, and inspection did not establish a separate concrete attack supporting a block.
Decision evidence
public snapshot- The CLI can read a saved Para vault token and uses BODIFY_URL, defaulting to Bodify, as the request destination.
- The CLI sends the token as a bearer credential, and its main command invokes the secrets lookup before continuing.
- The CLI describes the vault lookup as a way to pull secrets and tells users to connect the vault in Para Settings.
- The flagged CLI location concerns Claude worktree paths; the flagged declaration describes child-agent definitions, and the runtime reads project instruction files.
- The manifest has prepare and prepublishOnly scripts but no install hook; prepare configures a Git hooks path.
- The secrets lookup returns early if no token is available and fetches environment values from a Bodify API route.
- The inspected agent integration locations describe worktree handling, child-agent declarations, and project instruction loading; they do not establish an agent configuration takeover.
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@0.99.101.
See all versions of @livx.cc/agentx for other recorded verdicts.
- Review the evidence and your use of @livx.cc/agentx@0.99.101 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.