instar@1.3.1353
Coherence infrastructure for self-evolving AI agents — on the Claude Code or Codex subscription you already have.
AI Security Review
scanned 10d ago · by lpm-firewall-aiThe source shows a concrete, default-on agent hook auto-arming capability during runtime migration, so this warrants a warning under the agent-extension lifecycle policy. The inspected behavior is package-aligned and does not establish a malicious attack.
Decision evidence
public snapshot- When server startup detects an installed-version change, it runs the post-update migrator. The npm postinstall is a separate SQLite repair script.
- If codex-cli is enabled, the migrator registers hooks in the project and attempts to arm them; auto-arming is on by default and can be opted out in config.
- The PostUpdateMigrator target location describes framework-specific migration gating for agent files.
- SessionManager’s MCP configuration behavior is behind an explicit enable flag.
- The EvolutionManager, Config and ProjectMapper target locations inspect agent-related directories, binary paths or project filenames.
- The CapabilityMapper and CapabilityRegistryGenerator targets read agent skills or settings, while the AgentServer target reads Claude project transcripts.
- The Codex hook branch is conditional on codex-cli being enabled and installs Instar enforcement hooks.
- The observed capability is package-aligned agent setup; the inspected source does not establish credential theft or another malicious attack.
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@1.3.1353.
See all versions of instar for other recorded verdicts.
- Review the evidence and your use of instar@1.3.1353 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.