@vatio-ai/cli@0.104.0
Vatio CLI — deploy and manage Vatio agent workspaces
AI Security Review
scanned 10d ago · by lpm-firewall-aiSource inspection found no npm install lifecycle hook or concrete attack behavior. Agent guidance is created by the explicit init command, while saved-token requests use the configured service and transcript sharing requires affirmative user action.
Decision evidence
public snapshot- The package exposes a CLI entry point and declares yaml as its dependency; no package install hook is present.
- The user-invoked init command creates the workspace note; it leaves an existing Vatio section alone.
- Saved-token requests use the configured service URL. Session sharing requires confirmation or explicit --yes before sending a transcript.
- The skills hint is reached from a CLI command and checks the skills package version.
- The inspected dependency and utility paths implement YAML parsing, composition, and stringifying, plus ordinary CLI configuration and workspace functions; no install-triggered attack chain was identified.
- The CLI announces its command telemetry before sending it, and the telemetry endpoint follows the configured service URL.
- The init command writes coding-agent guidance into a project AGENTS.md, but only when the user explicitly runs init.
- The CLI can send session transcripts after confirmation and routine command telemetry during normal CLI use.
Source & flagged code
6 flaggedManifest-trigger-reachable source writes behavior-bearing configuration into a user or project AI-agent control surface.
lib/skills-hint.mjsView on unpkgFindings
Affected versions and remediation
This report applies to @vatio-ai/cli@0.104.0.
See all versions of @vatio-ai/cli for other recorded verdicts.
- Review the evidence and your use of @vatio-ai/cli@0.104.0 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.