karajan-code@4.51.1
Local multi-agent coding orchestrator with TDD, SonarQube, and code review pipeline
AI Security Review
scanned 10d ago · by lpm-firewall-aiThe active postinstall hook makes an unprompted change to two AI-agent configurations by registering the package’s own local MCP server. This matches the warning-only policy for first-party agent extension setup; the inspected source does not establish credential theft or another concrete attack.
Decision evidence
public snapshot- package.json runs scripts/postinstall.js during npm installation.
- The hook writes karajan-mcp registrations into the user’s Claude Code and Codex configuration without a consent check.
- The registration starts this package’s local MCP server over stdio.
- Other inspected agent integrations are CLI-driven or read-only, which limits the confirmed concern to the automatic MCP setup.
- The registered command points to the package’s own local server; the inspected setup does not send credentials or data to a network endpoint.
- The separate ai-trash hook installer requires an explicit install command with --claude-code.
Source & flagged code
24 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.
scripts/dual-publish.mjsView on unpkg · L17Findings
Affected versions and remediation
This report applies to karajan-code@4.51.1.
See all versions of karajan-code for other recorded verdicts.
- Review the evidence and your use of karajan-code@4.51.1 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.