devez-vibe@1.9.60

Stable terminal UI for Codex and Claude Agent SDK

AI Security Review

scanned 9d ago · by lpm-firewall-ai

The automatic postinstall setup places package-owned skills in user-level agent environments and can prune files within matching skill directories. This fits the warning category for first-party agent extension setup; the inspected source does not establish a block-worthy attack.

Static reason

High-risk behavior combination matched malicious policy.; source matched previously finalized malicious package; routed for review; previous stored version diff introduced dangerous source

Trigger

Installing the package through npm runs its postinstall hook.

Impact

This adds package-owned capabilities to agent environments without an interactive prompt and can replace or remove files within matching skill directories. No confirmed credential transfer or remote execution was established.

Mechanism

The hook copies the bundled skill trees into user-level agent skill directories, then removes destination entries absent from the bundled trees.

AI rationale

Evidence

package.jsoninstall-skills.mjsbridge/claude-agent-sdk-bridge.mjsskills/luna-loop/skills/insane-search/

Decision evidence

public snapshot

AI called this Suspicious at 88.0% confidence as Dangerous Capability with medium false-positive risk.

Evidence for warning

package.json runs install-skills.mjs automatically during postinstall.
install-skills.mjs targets user-level Codex and Claude skill directories, with configurable home locations.
The installer copies each bundled skill into those directories and prunes files not present in the package copy.
The prune step removes destination entries that are absent from the corresponding bundled skill.
The bridge changes Claude permission settings only through a named permissions/update request and checks whether managed policy locks those changes.

Evidence against

The inspected bridge permission writer is request-driven and respects the managed-policy lock; no automatic permission change was established.

Behavioral surface

Source

ChildProcessCryptoEnvironmentVarsFilesystem

Supply chain

HighEntropyStringsUrlStrings

ManifestNo manifest risk signals triggered.

Source & flagged code

8 flagged

package.jsonView file

Lines 25-51json

25 "bin/dvz.js",

26 "bin/dvz.exe",

27 "bridge/claude-agent-sdk-bridge.mjs",

28 "install-skills.mjs",

29 "prepare-search-deps.mjs",

30 "skills/luna-loop/",

31 "skills/insane-search/",

32 "README.md",

33 "LICENSE"

34 ],

35 "os": [

36 "win32"

37 ],

38 "cpu": [

39 "x64"

40 ],

41 "engines": {

42 "node": ">=18"

43 },

44 "scripts": {

45 "postinstall": "node install-skills.mjs && node prepare-search-deps.mjs"

High

Install Time Lifecycle Scripts

Package defines install-time lifecycle scripts.

package.jsonView on unpkg · L45

Medium

Ambiguous Install Lifecycle Script

Install-time lifecycle script is not statically allowlisted and needs review.

package.jsonView on unpkg · L45

46 },

47 "dependencies": {

48 "@anthropic-ai/claude-agent-sdk": "0.3.296"

49 }

50}

install-skills.mjsView file

Lines 1-21javascript

1import {

2 copyFileSync,

3 existsSync,

4 mkdirSync,

5 readdirSync,

6 rmSync,

7 writeFileSync,

8} from "node:fs";

9import { homedir } from

bin/dvz.exeView file

•path = bin/dvz.exe kind = native_binary sizeBytes = 26684416 magicHex = [redacted]

Medium

Ships Native Binary

Package ships native binary artifacts.

bin/dvz.exeView on unpkg

skills/insane-search/tests/live_benchmark/bench.pyView file

•path = skills/insane-search/tests/live_benchmark/bench.py kind = payload_in_excluded_dir sizeBytes = 4268 magicHex = [redacted]

High

Payload In Excluded Dir

Package hides binary, compressed, or executable-looking payloads in test/fixture/hidden paths.

skills/insane-search/tests/live_benchmark/bench.pyView on unpkg

•path = skills/insane-search/tests/live_benchmark/bench.py kind = build_helper sizeBytes = 4268 magicHex = [redacted]

Medium

Ships Build Helper

Package ships non-JavaScript build or shell helper files.

skills/insane-search/tests/live_benchmark/bench.pyView on unpkg

bin/dvz.jsView file

•matchType = normalized_sha256 matchedPackage = devez-vibe@1.9.41 matchedPath = bin/dvz.js matchedIdentity = npm:ZGV2ZXotdmliZQ:1.9.41 similarity = 1.000 summary = normalized source hash matched finalized malicious source

High

Known Malware Source Similarity

Source file is highly similar to a previously finalized malicious package; route for source-aware review.

bin/dvz.jsView on unpkg

bridge/claude-agent-sdk-bridge.mjsView file

•matchType = previous_version_dangerous_delta matchedPackage = devez-vibe@1.9.52 matchedIdentity = npm:ZGV2ZXotdmliZQ:1.9.52 similarity = 0.833 summary = stored previous version shares package body but lacks this dangerous source file

High

Previous Version Dangerous Delta

This package version adds a dangerous source file absent from the previous stored version; route for source-aware review.

bridge/claude-agent-sdk-bridge.mjsView on unpkg

Source & flagged code

8 flagged · loading source

package.jsonView file

•scripts.postinstall = node install-skills.mjs && node prepare-search-deps.mjs

High

Install Time Lifecycle Scripts

Package defines install-time lifecycle scripts.

package.jsonView on unpkg

•scripts.postinstall = node install-skills.mjs && node prepare-search-deps.mjs

Medium

Ambiguous Install Lifecycle Script

Install-time lifecycle script is not statically allowlisted and needs review.

package.jsonView on unpkg

install-skills.mjsView file

1import { L2: copyFileSync, L3: existsSync, L4: mkdirSync, L5: readdirSync, L6: rmSync, L7: writeFileSync, L8: } from "node:fs"; ... L16: const userHome = homedir(); L17: const codexHome = process.env.CODEX_HOME?.trim() || join(userHome, ".codex"); L18: const claudeHome = process.env.CLAUDE_CONFIG_DIR?.trim() || join(userHome, ".claude"); L19:

Critical

Ai Agent Control Hijack

Source creates an unconsented AI-agent control surface through install-time mutation or a default unauthenticated remote skill channel.

Findings

1 Critical4 High5 Medium4 Low

CriticalAi Agent Control Hijackinstall-skills.mjs

HighInstall Time Lifecycle Scriptspackage.json

HighPayload In Excluded Dirskills/insane-search/tests/live_benchmark/bench.py

HighKnown Malware Source Similaritybin/dvz.js

HighPrevious Version Dangerous Deltabridge/claude-agent-sdk-bridge.mjs

MediumAmbiguous Install Lifecycle Scriptpackage.json

MediumEnvironment Vars

MediumShips Native Binarybin/dvz.exe

MediumShips Build Helperskills/insane-search/tests/live_benchmark/bench.py

MediumStructural Risk Force Deep Review

LowScripts Present

LowFilesystem

Affected versions and remediation

This report applies to devez-vibe@1.9.60.

See all versions of devez-vibe for other recorded verdicts.

Review the evidence and your use of devez-vibe@1.9.60 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: 2026-10-13 23:46:04.936Z (UTC).

Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.