@su-record/vibe@4.1.15
An AX/FDE harness for Claude Code, Codex CLI and ChatGPT desktop. Say what you need, approve the scenarios once; the harness proves the work by running the checks itself. Memory in plain files, human tokens for irreversible actions, a ledger instead of cl
AI Security Review
scanned 11d ago · by lpm-firewall-aiSource inspection confirms a guarded global-install setup that can modify detected AI client homes with Vibe-owned surfaces. This warrants a warning under the agent integration policy, but the inspected behavior does not establish malicious activity.
Decision evidence
public snapshot- package.json defines a postinstall script that runs scripts/postinstall.js.
- The postinstall setup runs only for global installs and can be skipped with VIBE_SKIP_SETUP; it imports and calls setupGlobal.
- setupGlobal installs surfaces for detected client homes, and installPlugin assembles and registers the package's plugin.
- register.js reads the Claude plugin's installed version and marketplace configuration.
- The inspected GitHub client sends authenticated requests to api.github.com, its configured GitHub API host.
- The check runner executes scenario commands in the project root, consistent with the package's verification harness.
Source & flagged code
16 flagged · loading sourcePackage defines install-time lifecycle scripts.
package.jsonView on unpkgInstall-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgPackage source references child process execution.
mcpb/server/index.jsView on unpkg · L10Findings
Affected versions and remediation
This report applies to @su-record/vibe@4.1.15.
See all versions of @su-record/vibe for other recorded verdicts.
- Review the evidence and your use of @su-record/vibe@4.1.15 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.