sneakoscope@10.6.2
Sneakoscope Codex (`sks`) is an open-source trust layer for Codex CLI and ChatGPT Desktop. It coordinates bounded AI coding agents, records machine-verifiable evidence, preserves project memory, and blocks release claims that are not supported by current
AI Security Review
scanned 3d ago · by lpm-firewall-aiThe postinstall hook can perform first-party Menu Bar setup before the separate bootstrap opt-in, including creating a persistent launch agent. The inspected runtime receiver parses provider events for compaction handling; no malicious behavior was established there.
Decision evidence
public snapshot- The package postinstall command invokes the installation helper.
- Postinstall attempts Menu Bar setup before checking the separate bootstrap opt-in.
- Automatic Menu Bar setup is limited to interactive global installs on macOS and runs the installer with apply enabled.
- The installer creates a package-owned launch agent configured to run at load and stay alive.
- The flagged runtime receiver parses provider event data, and its caller uses the parsed result as a compaction summary.
- The broader postinstall bootstrap requires an affirmative opt-in.
- The Menu Bar setup is gated by platform, global-install, CI, and terminal checks; no credential transfer or malicious action was confirmed.
Source & flagged code
14 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 contains a possible secret pattern.
dist/core/codex/managed-proxy-env.jsView on unpkg · L43Findings
Affected versions and remediation
This report applies to sneakoscope@10.6.2.
See all versions of sneakoscope for other recorded verdicts.
- Review the evidence and your use of sneakoscope@10.6.2 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.