@vatio-ai/cli@0.75.0
Vatio CLI — deploy and manage Vatio agent workspaces
AI Security Review
scanned 25d ago · by lpm-firewall-aiThe package has no install lifecycle hook or confirmed malicious behavior. Its explicit `init` command changes an agent instruction file, a bounded agent capability that warrants a warning under the review policy.
Static reason
No blocking static signals were detected.
Trigger
A user runs `vatio init` for a target directory.
Impact
Coding agents reading that file may follow the added Vatio workflow guidance.
Mechanism
The command creates `AGENTS.md` or appends the Vatio section when the existing file does not already mention Vatio.
AI rationale
The package has no install lifecycle hook or confirmed malicious behavior. Its explicit `init` command changes an agent instruction file, a bounded agent capability that warrants a warning under the review policy.
Evidence
lib/commands/misc.mjsAGENTS.md
Decision evidence
public snapshotAI called this Suspicious at 87.0% confidence as Dangerous Capability with low false-positive risk.
Evidence for warning
- The package exposes a CLI and declares no npm lifecycle hooks.
- Running `vatio init` writes a Vatio section into the target directory’s `AGENTS.md`.
- The added agent guidance directs work through Vatio commands and workspace documentation.
- The skills hint is gated on selected CLI commands and can be disabled in CI or with an environment flag.
- Session sharing asks before sending the conversation and supports saving it for review first.
Evidence against
- The `init` command modifies an AI-agent instruction file, so it has a concrete, user-invoked agent configuration effect.
Behavioral surface
ChildProcessCryptoEnvironmentVarsFilesystemNetwork
HighEntropyStringsUrlStrings
NoLicense
Source & flagged code
6 flaggedlib/skills-hint.mjsView file
•Manifest-trigger-reachable source links an external AI-agent control path to a behavior-bearing write operation.
import { existssync, mkdirsync, readfilesync, writefilesync } from "node:fs";
import { homedir } from "node:os";
import { dirname, join, resolve } from "node:path";
import { confighome } from "./config.mjs";
import { paint } from "./support.mjs";
const skill = "vatio-improve";
const version_file = ".vatio-sk[redacted]";
const latest_url = "https://registry.npmjs.org/@vatio-ai/skills/latest";
const show_every_ms = 24 * 60 * 60 * 1000;
const check_every_ms = 24 * 60 * 60 * 1000;
const fetch_timeout_ms = 1_500;
export async function printskillshint(startdir = process.cwd()) {
if (process.env.ci || process.env.vatio_no_skills_hint) return;
const path = join(confighome(), "sk
High
Trigger Reachable External Ai Agent Control Surface Mutation
Manifest-trigger-reachable source writes behavior-bearing configuration into a user or project AI-agent control surface.
lib/skills-hint.mjsView on unpkg32export async function printSkillsHint(startDir = process.cwd()) {
if (process.env.CI || process.env.VATIO_NO_SKILLS_HINT) return;
const path = join(configHome(), "sk[redacted]");
const key = resolve(startDir);
const s
Medium
Ai Review Evidence
The skills hint is gated on selected CLI commands and can be disabled in CI or with an environment flag.
lib/skills-hint.mjsView on unpkg · L32Findings
2 High6 Medium4 Low
HighTrigger Reachable External Ai Agent Control Surface Mutationlib/skills-hint.mjs
HighSemantic Analysis Limitedlib/commands/support-commands.mjs#virtual:normalized:round1
MediumNetwork
MediumEnvironment Vars
MediumSuspicious Lifecycle Evidencepackage.json
MediumAi Review Evidencelib/commands/misc.mjs
MediumAi Review Evidencelib/commands/misc.mjs
MediumAi Review Evidencelib/skills-hint.mjs
LowFilesystem
LowHigh Entropy Strings
LowUrl Strings
LowNo License
Affected versions and remediation
This report applies to @vatio-ai/cli@0.75.0.
See all versions of @vatio-ai/cli for other recorded verdicts.
- Review the evidence and your use of @vatio-ai/cli@0.75.0 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.