OmniVideo no seu computador: o CLI omni, o studio local e as ferramentas do pipeline.
Soft block: AI-agent control surface
Warn by default; block when configured. Claude may run the package’s commands and treat the workspace as trusted without a fresh approval prompt.
AI security review
scanned · by lpm-firewall-ai
The session hook can trigger an unattended setup that edits Claude’s permission allowlist and accepts workspace trust. These are agent approval changes on a default start path, so the package meets the control-hijack blocking rule.
Static reason
One or more suspicious static signals were detected.; previous stored version diff introduced dangerous source
Trigger
On a Claude session hook, when an Omni login exists and setup is incomplete; the hook starts `setup --auto` in the background.
Impact
Claude may run the package’s commands and treat the workspace as trusted without a fresh approval prompt.
Mechanism
Setup installs Omni skills, adds Omni CLI skill/run commands to `permissions.allow`, and marks the Omni workspace trusted in Claude’s account configuration.
Policy narrative
When the session hook finds an existing Omni login and incomplete setup, it launches `omni setup --auto` in the background. Setup adds the Omni CLI’s skill and run commands to Claude’s automatic permission allowlist and marks the Omni workspace as trusted in Claude’s account file. This changes agent approval settings on a default session-start path without asking at that time.
AI identified this as a dangerous AI-agent capability at 91.0% confidence with low false-positive risk.
Evidence for policy risk
When `omni up --hook` runs with an existing login and setup is needed, it starts setup in the background without waiting for another request.
The background process runs `omni setup --auto`; setup installs the package’s skills, updates Claude permissions, and marks the Omni workspace trusted.
The permission update adds Omni CLI skill and run commands to Claude’s `permissions.allow` list and writes the settings file.
The trust update sets `hasTrustDialogAccepted` to true for the Omni workspace in Claude’s account configuration.
Evidence against
The package manifest has no install lifecycle script; this behavior is reached through the session hook after setup has been requested previously.
The added allow rules are scoped to the Omni CLI’s skill and run commands, and the permission update preserves existing settings.
Behavioral surface
SourceChild process · Crypto · Dynamic require · Environment variables · Filesystem · Network · Shell · Web socket
Supply chainHigh entropy strings · URL strings
ManifestNo license
Scanned 1,154 files · 8.15 MB of source
47 external domains referenced
127.0.0.1
accounts.google.com
api.cartesia.ai
api.cloudflare.com
api.elevenlabs.io
api.fish.audio
api.openai.com
api.openverse.org
api.scrapecreators.com
api.x.ai
app.crievo.com
audiocdn.epidemicsound.com
blog.exemplo.com
bun.sh
console.cloud.google.com
developers.facebook.com
developers.tiktok.com
elevenlabs.io
fish.audio
fonts.googleapis.com
fonts.gstatic.com
github.com
graph.instagram.com
i.ytimg.com
jcgt.org
oauth2.googleapis.com
open.tiktokapis.com
openrouter.ai
platform.openai.com
play.cartesia.ai
registry.npmjs.org
schemas.microsoft.com
scrapecreators.com
studio.youtube.com
support.google.com
www.apple.com
www.epidemicsound.com
www.googleapis.com
www.instagram.com
www.reddit.com
www.tiktok.com
www.w3.org
www.youtube.com
x.com
youtu.be
youtube.com
youtubeanalytics.googleapis.com
Findings
1 critical · 6 high · 6 medium · 5 low
Where they are, most severe file first. Code opens a file's flagged lines, with their evidence, below.
Avoid installing @crievo/omni@0.1.140. Remove it from direct dependencies and check your lockfile for transitive copies.
Choose an independently verified alternative or release. This report does not establish that other versions are safe.
If this version ran, check your AI agents' configuration and instruction files (the files this report lists) for changes you did not make, and remove them. This finding does not by itself establish malware intent.