@makerx/yoke@0.7.0
Lean, user-invoked Claude Code skills for MakerX AI-assisted delivery, plus the CLI they call.
AI Security Review
scanned 3d ago · by lpm-firewall-aiThe package has no install-time execution hook, and the CLI's process, network, and file operations are tied to user-invoked review, setup, and configuration commands. The scanner's command-output and agent-configuration signals match those documented CLI functions; the inspected code does not show automatic exfiltration or an unauthorized default write.
Decision evidence
public snapshot- The manifest exposes a CLI and has only a `prepublishOnly` build script; no install hook runs on registry install.
- The `.claude/skills` writes occur only in the explicit `yoke sync` command, with user-scope writes requiring `--user`.
- The review command is user-invoked; provider probes send a `ping` to the configured provider endpoint using its configured key.
- The CLI can execute `gh`, `git`, Claude, and Codex processes and can access their output as part of its requested review and PR workflows.
- Review operations can contact GitHub or a configured model provider, and `review --post` can publish findings when explicitly requested.
Source & flagged code
4 flaggedSource combines command execution, command-output handling, and outbound requests; review data flow before blocking.
dist/cli.jsView on unpkg · L1223Findings
Affected versions and remediation
This report applies to @makerx/yoke@0.7.0.
See all versions of @makerx/yoke for other recorded verdicts.
- Review the evidence and your use of @makerx/yoke@0.7.0 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.