registry  /  ttj-skills-browser  /  1.2.6

ttj-skills-browser@1.2.6

ttj-skills-browser - Playwright CLI with auto-profile management

AI Security Review

scanned 4h ago · by lpm-firewall-ai

LPM blocks this version under the AI-agent control-surface policy. Installation mutates global Claude Code and Codex agent-skill directories. This occurs automatically during npm postinstall, before the user invokes the CLI.

Static reason
High-risk behavior combination matched malicious policy.
Trigger
`npm install` executes the package postinstall lifecycle hook.
Impact
Adds package-controlled instructions to broad user-level AI-agent control surfaces.
Mechanism
Unconsented global AI-agent skill installation.
Policy narrative
On installation, npm invokes `scripts/install-skill.js`. The script copies the package-supplied skill into the user's global Claude Code skill directory and writes a converted version into the global Codex agent skill directory, then writes an installation marker. These are foreign, broad AI-agent control surfaces changed without an explicit user command.
Rationale
The package contains a concrete unconsented `postinstall` mutation of global AI-agent instruction directories. Under the install-control-surface policy, this warrants blocking even though no credential theft or exfiltration was found.
Evidence
package.jsonscripts/install-skill.js.claude/skills/ttj-skills-browser/SKILL.md~/.claude/skills/ttj-skills-browser/SKILL.md~/.agents/skills/ttj-skills-browser/SKILL.md~/.ttj-skills-browser-installed

Decision evidence

public snapshot
AI called this Malicious at 97.0% confidence as Malware with low false-positive risk.
Evidence for policy block
  • `package.json` runs `postinstall: node scripts/install-skill.js`.
  • `scripts/install-skill.js` unconditionally copies packaged `SKILL.md` to `~/.claude/skills/ttj-skills-browser/SKILL.md`.
  • The same postinstall writes a converted skill to `~/.agents/skills/ttj-skills-browser/SKILL.md`.
  • Postinstall also creates `~/.ttj-skills-browser-installed` without user invocation.
Evidence against
  • Installer does not read credentials or transmit data.
  • No obfuscated payload, dynamic code loading, or remote execution was found.
  • Runtime network use is package-aligned: npm version lookup and browser launch.
Behavioral surface
Source
ChildProcessEnvironmentVarsFilesystemNetworkShell
Supply chain
HighEntropyStringsUrlStrings
ManifestNo manifest risk signals triggered.
scanned 7 file(s), 37.7 KB of source, external domains: 127.0.0.1, registry.npmjs.org, www.google.com

Source & flagged code

3 flagged · loading source
package.jsonView file
scripts.postinstall = node scripts/install-skill.js
High
Install Time Lifecycle Scripts

Package defines install-time lifecycle scripts.

package.jsonView on unpkg
dist/utils.jsView file
5import path from 'path'; L6: import net from 'net'; L7: import https from 'https'; L8: import { exec, execFile } from 'child_process'; L9: import { promisify } from 'util'; ... L19: */ L20: export const getOsType = () => process.platform === 'win32' L21: ? 'windows' ... L27: */ L28: export const getHomeDir = () => os.homedir(); L29: /** ... L35: const isWindows = getOsType() === 'windows';
High
Sandbox Evasion Gated Capability

Source gates dangerous network, credential, or execution behavior behind CI, host, platform, time, or geo fingerprint checks.

dist/utils.jsView on unpkg · L5
scripts/install-skill.jsView file
14Install-time AI-agent control hijack evidence: L14: const skillName = 'ttj-sk[redacted]'; L15: const sk[redacted] = path.join(__dirname, '..', '.claude', 'skills', skillName); L16: const skillFile = 'SKILL.md'; ... L19: if (!fs.existsSync(dir)) { L20: fs.mkdirSync(dir, { recursive: true }); L21: } ... L43: // Claude Code: 기존 /ttj-sk[redacted] 호출 방식 유지 L44: const claudeTargetDir = path.join(homedir(), '.claude', 'skills', skillName); L45: const claudeTargetFile = path.join(claudeTargetDir, skillFile); L46: ensureDir(claudeTargetDir); L47: fs.copyFileSync(sourceFile, claudeTargetFile); L48: console.log(`✅ Claude Code 스킬 설치 완료: ${claudeTargetFile}`); Payload evidence from .claude/skills/ttj-sk[redacted]/SKILL.md: L133: L134: **HTTP 요청**: Axios + Cheerio 조합 L135: - Axios: HTTP 요청 수행 ... L270: L271: ```powershell L272: # Windows PowerShell
Critical
Ai Agent Control Hijack

Install-time source drops package-supplied AI-agent/MCP control files or instructions.

scripts/install-skill.jsView on unpkg · L14

Findings

1 Critical2 High3 Medium5 Low
CriticalAi Agent Control Hijackscripts/install-skill.js
HighInstall Time Lifecycle Scriptspackage.json
HighSandbox Evasion Gated Capabilitydist/utils.js
MediumNetwork
MediumEnvironment Vars
MediumStructural Risk Force Deep Review
LowNon Install Lifecycle Scripts
LowScripts Present
LowFilesystem
LowHigh Entropy Strings
LowUrl Strings