Back to skill

Security audit

🎬 观影小管家

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent movie assistant, but it mishandles media-server credentials and stores personal viewing context with weak controls.

Install only after reviewing and fixing credential handling: remove hard-coded API keys, rotate exposed keys, require user-provided secrets, separate public API requests from Emby-authenticated requests, avoid sending Emby tokens in URLs, and prefer HTTPS. Users should also understand that the skill keeps local records of viewing history, ratings, mood, work status, and interests, and should have clear ways to review, delete, or disable that memory.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (3)

T09 · Insecure Skill Coding Practices

Error
Location
index.js:29
Finding

Emby API Token Disclosed to Unrelated TMDB Endpoints

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
index.js:15
Finding

API Credentials Hard-Coded in Source Code and Documentation

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
index.js:113
Finding

Emby Credential Exposed in Query Strings over a Default Plaintext Connection

Content
View full analysis
{ try { const res = await fetchJson(`${EMBY_URL}/Items/${id}?api_key=${EMBY_API_KEY}`); return { name: res.Name, type: res.CollectionType }; ``` ### Technical Analysis The Emby API token is placed in the URL query string even though the shared HTTP helper also sends it in the `X-Emby-Token` header. Query-string credentials can be retained in server access logs, reverse-proxy logs, monitoring systems, browser or debugging histories, and error reports. The default Emby URL uses unencrypted HTTP. When this default or another HTTP endpoint is used, both authentication material and private media-library traffic are transmitted without transport confidentiality or server authentication. A network-positioned attacker could observe the request or modify its response. Sending the token in both the header and URL is unnecessary. The query-string copy exceeds the minimum authentication data needed for the declared Emby functionality. ### Attack Path 1. A user configures an Emby API token but leaves the default HTTP server URL or configures an ...[truncated 1281 chars]
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • YARA SignaturesMalware Match, Webshell Match, Cryptominer Match
Findings (34)

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 101)May include surrounding context.

md
- `SKILL.md` - 本说明文档

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 102)May include surrounding context.

md
- `index.js` - 主程序

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · feishu-card.js (reported line 9)May include surrounding context.

js
*/

const path = require('path');
require('dotenv').config({ path: path.join(__dirname, '../../../.env') });

const EMBY_URL = process.env.EMBY_URL || 'http://192.168.0.151:8096';

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · index.js (reported line 12)May include surrounding context.

js
*/

const path = require('path');
require('dotenv').config({ path: path.join(__dirname, '../../../.env') });

const EMBY_URL = process.env.EMBY_URL || 'http://192.168.0.151:8096';

YARA rule 'agent_skill_mcp_tool_poisoning_metadata': MCP/tool metadata poisoning indicators in tool schemas or skill manifests [agent_skills]

High
Category
YARA Match
Confidence
80% confidence
Finding

YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).

Content

Scanner excerpt · package.json (reported line 5)May include surrounding context.

json
{
  "name": "movie-butler",
  "displayName": "🎬 观影小管家",
  "version": "1.0.0",
  "description": "杜老师的专属观影助手,整合 TMDB + Emby,提供电影查询、媒体库管理、个性化推荐、每周观影计划等服务",
  "author": "小暖阳",
  "license": "MIT",
  "tags": ["电影", "Emby", "TMDB", "推荐", "媒体库"],
  "category": "娱乐",
  "requirements": {
    "env": [
      "TMDB_API_KEY",
      "EMBY_URL",
      "EMBY_API_KEY",
      "EMBY_USER_ID"
    ]
  },
  "features": [
    "🌐 双源查询(TMDB + Emby)",
    "🔗 一键播放链接",
    "📊 精确年份匹配",
    "🎯 智能推荐算法",
    "👨‍👩‍👦 家庭电�

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill explicitly states it will record the user's mood, work status, interests, and later use that information for recommendations, but provides no notice about storage, retention, sharing, or user control. This creates a privacy risk because it encourages collection of behavioral and potentially sensitive personal data without informed consent or clear limits.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The demo shows persistent storage of viewing history, ratings, and subjective feedback, but does not warn users that these become lasting personal records. Viewing habits and ratings can reveal preferences, routines, and potentially sensitive inferences, so silent persistence increases privacy and trust risks.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

Referencing an Emby API key without warning about the scope of access understates the privacy impact of connecting to a media server. An API key may allow broad access to server content and metadata, so users should be informed that the skill may inspect their library and potentially infer household viewing behavior.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill advertises recording movie impressions, ratings, and background information, but it does not prominently warn users that these preferences and status signals will be persistently stored. Because the stored data includes mood, work state, interests, and media habits, inadequate disclosure creates a meaningful privacy and consent issue.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
99% confidence
Finding

The README discloses a real TMDB API key and internal Emby server details including a private IP address, username, and service endpoint. Even if some values are incomplete, publishing live-looking credentials and internal infrastructure information can enable unauthorized API usage, service abuse, reconnaissance, and targeted attacks against the media server environment.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The example trigger phrases are very broad natural-language commands that can easily overlap with ordinary conversation, causing the skill to activate unintentionally. In a skill that queries external services and may reveal server availability or generate downstream actions like download suggestions, accidental invocation increases privacy and safety risk.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The background-recording examples use everyday statements like mood and work status without clear boundaries indicating that these will be stored as persistent memory. This can cause users to unknowingly save sensitive personal context, creating privacy risk and potentially influencing future recommendations in ways they did not intend.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

Examples like '本周推荐什么电影?', '这周有什么好电影推荐', and similar everyday phrases are broad enough to be said in normal conversation and may invoke the skill unintentionally. Because the skill can query external services and access personal media-server data, accidental activation could disclose preferences or trigger unwanted network requests.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The background-recording examples are ordinary personal statements like mood or work status, but the documentation frames them as inputs that improve recommendations without strong opt-in boundaries. This creates risk that sensitive behavioral data is captured or persisted when the user does not clearly understand they are authorizing storage and profiling.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The file structure notes that 'movie-memory.md' is auto-created for viewing records and preferences, but the skill does not clearly warn users that reactions, ratings, and inferred tastes will be stored persistently. Persistent storage of entertainment history and emotional reactions can reveal personal preferences and habits, especially on shared systems.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
99% confidence
Finding

The documentation exposes a real-looking TMDB API key and internal service configuration details, including a private LAN Emby URL and username. Hard-coded secrets and infrastructure identifiers can enable unauthorized API use, credential stuffing against related services, or reconnaissance of the user's environment, and this is unrelated to what end users need to invoke the skill safely.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The documentation describes configuration for TMDB, Emby, and Plex but does not clearly tell users that their search queries and media-related requests may be sent to those third-party or self-hosted services using configured credentials. That omission weakens informed consent and can surprise users who may not expect external transmission of interests, watch history context, or server queries.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · index.js (reported line 16)May include surrounding context.

js
// 配置
const TMDB_API_KEY = process.env.TMDB_API_KEY || 'bd1ba3aa647fbaa7b35e93db5164a53f';
const TMDB_BASE_URL = 'https://api.themoviedb.org/3';
const OMDb_API_KEY = process.env.OMDB_API_KEY || '9e58a5e3'; // OMDb API (IMDb 数据)
const EMBY_URL = process.env.EMBY_URL || 'http://192.168.0.151:8096';
const EMBY_USERNAME = process.env.EMBY_USERNAME || '喜悦影音';

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The inline documentation says the skill will only search movie and series libraries, but the subsequent Emby code explicitly includes musicvideos folders and MusicVideo item types in multiple queries. This is an active contradiction between stated intent in comments and implemented scope.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The skill hardcodes zh-CN in external API requests and uses only Chinese trigger terms, examples, and output text throughout the file. This enforces a specific language/locale without any opt-in or alternate language selection, which matches the stated policy-violation category.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The function comment states it retrieves specified libraries for movies, series, and concerts, and the following comment narrows this to 'movie, series, concert' libraries, but the implemented filter actually uses Emby's musicvideos collection type. That behavior is materially different from the natural-language description and broadens the content category accessed.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The code reads Emby watch history and persists it into a local markdown file without any consent, retention control, or disclosure in the code flow. Viewing history is personal behavioral data, and silently storing it increases privacy risk if the host is multi-user, backed up elsewhere, or later exposed through logs or file access.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill saves user ratings and free-form notes to disk without notifying the user that the information will be persisted locally. Because the notes can contain personal opinions or sensitive details, undisclosed storage creates a privacy issue and can lead to unintended disclosure on shared systems.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

This code stores sensitive personal context such as mood, work status, and interests to a local file without explicit warning or consent. Such contextual data is more privacy-sensitive than movie metadata and could be harmful if accessed by other local users, backup systems, or later-integrated tooling.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

All user-facing instructions and examples are presented only in Chinese, and the skill does not offer an alternative language or indicate that Chinese is an intentional region-specific requirement. Under the policy, forcing a specific language without user opt-in is a natural-language policy concern.

Content

No source excerpt is available for this finding.