registry / @varveai/adit-cli / 1.0.9
@varveai/adit-cli@1.0.9
AI Development Intent Tracker — Transparent Time Machine for your codebase
AI Security Review
scanned 5d ago · by lpm-firewall-aiThe suspicious scanner hits resolve to expected cloud-agent hook setup and user-invoked update behavior. The package has no install hook, and the reviewed hook endpoint is loopback rather than an unrelated collector.
Static reason
One or more suspicious static signals were detected.; previous stored version diff introduced dangerous source
Trigger
The Claude hooks are installed during an explicit `adit cloud claude` session; self-update runs only when the upgrade command is invoked.
Impact
No unauthorized credential collection, unrelated data transfer, or concealed automatic action was found.
Mechanism
The Claude status-line hook posts limited session metadata to a temporary loopback server. Cloud credentials are used for the configured ADIT cloud service after login.
AI rationale
The suspicious scanner hits resolve to expected cloud-agent hook setup and user-invoked update behavior. The package has no install hook, and the reviewed hook endpoint is loopback rather than an unrelated collector.
Evidence
.claude/adit-cloud-<terminalId>.settings.local.json~/.adit/cloud-credentials.json
Network endpoints1
127.0.0.1:<ephemeral-port>/hook
Decision evidence
public snapshotAI called this Clean at 92.0% confidence as Benign with low false-positive risk.
Evidence for AI clean decision
- The manifest declares only build and typecheck scripts; it has no install lifecycle hook.
- The cloud Claude command requires an initialized project and saved cloud credentials before starting.
- That command creates a loopback hook server and writes a session-specific settings file that points to it.
- The status-line hook sends model, effort, session, transcript-path, and working-directory metadata to the loopback endpoint.
- Cloud login is an explicit command, saves credentials locally, and calls an upload disclosure.
Evidence against
- The flagged shell execution is in the explicit self-update command, which runs package-manager installation as part of upgrading ADIT.
- The Codex resume watcher reads local Codex databases to detect a resume event; no outbound transfer is shown in that path.
Behavioral surface
ChildProcessCryptoEnvironmentVarsFilesystemNetworkShellWebSocket
Source & flagged code
8 flaggeddist/commands/self-update.jsView file
8*/
L9: import { execSync } from "node:child_process";
L10: import { existsSync, readFileSync } from "node:fs";
High
Child Process
Package source references child process execution.
dist/commands/self-update.jsView on unpkg · L8196console.log("Installing dependencies ...");
L197: execSync("pnpm install --frozen-lockfile", {
L198: ...showExecOpts,
High
Runtime Package Install
Package source invokes a package manager install command at runtime.
dist/commands/self-update.jsView on unpkg · L196dist/commands/cli-agent/codex/codex-resume-log-watcher.jsView file
184if (process.platform === "win32") {
L185: const { stdout } = await run("powershell.exe", [
L186: "-NoProfile", "-NonInteractive", "-Command",
Findings
1 Critical7 High3 Medium5 Low
CriticalPrevious Version Dangerous Deltadist/commands/cli-agent/claude/hooks-bootstrap.js
HighChild Processdist/commands/self-update.js
HighShelldist/commands/cli-agent/codex/codex-resume-log-watcher.js
HighSame File Env Network Executiondist/commands/cli-agent/claude/hooks-bootstrap.js
HighCommand Output Exfiltrationdist/commands/cli-agent/claude/hooks-bootstrap.js
HighTrigger Reachable Command Output Exfiltrationdist/commands/cli-agent/claude/hooks-bootstrap.js
HighRuntime Package Installdist/commands/self-update.js
HighSemantic Analysis Limitedpackage.json
MediumNetwork
MediumEnvironment Vars
MediumStructural Risk Force Deep Review
Affected versions and remediation
This report applies to @varveai/adit-cli@1.0.9.
See all versions of @varveai/adit-cli for other recorded verdicts.
- Review the evidence and your use of @varveai/adit-cli@1.0.9 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.