instar@1.3.1342
Coherence infrastructure for self-evolving AI agents — on the Claude Code or Codex subscription you already have.
AI Security Review
scanned 14d ago · by lpm-firewall-aiThe inspected agent configuration changes are tied to first-party setup and migration behavior, and the postinstall script repairs a native dependency. No concrete malicious behavior or unconsented foreign control-surface mutation was established, so this merits a warning for agent integration lifecycle risk rather than a block.
Decision evidence
public snapshot- The postinstall hook runs a native dependency repair script; the manifest shows no automatic hook that invokes the agent configuration migrator.
- The post-update migrator writes an Instar-owned hook into the agent's hook directory.
- The init command scaffolds Claude settings only when Claude Code is enabled.
- The migration command writes a session hook under the project’s Instar hook directory.
- Other inspected agent integration locations load package-owned skills or session helpers, or manage memory under the agent’s own configuration home.
- The flagged source behavior is tied to package-owned agent setup and an update migration, rather than an observed unconsented install-time change to a foreign agent configuration.
- No inspected excerpt establishes credential theft, destructive behavior, stealth persistence, or remote payload execution.
Source & flagged code
18 flagged · loading sourcePackage defines install-time lifecycle scripts.
package.jsonView on unpkgPackage source references child process execution.
dist/threadline/PipeSessionSpawner.jsView on unpkg · L15Package source references shell execution.
Findings
Affected versions and remediation
This report applies to instar@1.3.1342.
See all versions of instar for other recorded verdicts.
- Review the evidence and your use of instar@1.3.1342 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.