instar@1.3.1412
⚠ Under reviewCoherence infrastructure for self-evolving AI agents — on the Claude Code or Codex subscription you already have.
AI Security Review
scanned 1d ago · by lpm-firewall-aiThe inspected high-risk paths did not substantiate the scanner’s credential-redirection or persistence claims. A clean verdict requires citations for 419 paths, beyond the allowed 256, so the package needs further review.
Decision evidence
public snapshot- The manifest’s install hook runs a helper for the better-sqlite3 native binary; the helper describes testing or building that binary.
- The inspected Telegram request uses Telegram’s API host.
- The credential provider reads credentials from the operating system keychain and writes Claude OAuth fields back to that store.
- The inspected Git signing code writes an allowed-signers file under the package state directory.
- The inspected source excerpts did not establish the flagged credential redirect or persistence backdoor.
- A full clean review would require citations for 419 distinct paths, exceeding the 256 citation limit.
Source & flagged code
17 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@1.3.1412.
See all versions of instar for other recorded verdicts.
- Review the evidence and your use of instar@1.3.1412 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.