Back to skill

Security audit

Yeelight WiFi LAN Control

Security checks for vulnerabilities and agentic risk

Overview

This skill locally controls Yeelight WiFi lights and stores local home settings in a way that is disclosed and aligned with its purpose.

Install this only if you want a local agent to discover and control Yeelight WiFi lights on your LAN. Review prompts before confirming writes, schedules, recovery, music mode, or store reset, and be aware that some responses and confirmation text may appear in Chinese.

Vulnerability Patterns
  • 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
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Memory PoisoningPersistent Context Injection, Context Window Stuffing, Memory Manipulation
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (19)

Memory Manipulation

High
Category
Memory Poisoning
Confidence
85% confidence
Finding

Skill manipulates agent memory, state, or stored context. Memory corruption can alter personality, override safety rules, or cause unpredictable behavior.

Content

Scanner excerpt · SKILL.md (reported line 93)May include surrounding context.

md
Multi-device scene application is not transactional. The runtime persists a
pre-state/recovery record before writing, verifies every device, and reports exact
success, skipped, failed, or uncertain rows. Offer `operation.recover` only after
explicit confirmation. Recovery re-reads first and refuses to overwrite state that
has drifted outside the known transition.

## Scheduling

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
80% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 22)May include surrounding context.

printf '%s' '{"operation":"discover"}' | sh scripts/invoke.sh

text

PowerShell uses `scripts/invoke.ps1` with the same stdin/stdout contract. Never ask
the user to paste protocol commands, IP addresses, raw JSON, credentials, or
network packets. Never call `yeelight-home`, cloud APIs, MCP, OAuth, or a Pro home.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

The default prompt is broad enough to trigger this skill for general requests about discovering, organizing, or controlling lights without strong boundaries around user intent. Because this skill performs local network device discovery and state-changing actions on physical devices, unintended activation could cause unauthorized or surprising changes in a user's environment.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

Enabling implicit invocation without clear activation boundaries increases the chance that the skill will run automatically on ambiguous home-automation requests. In this context, the skill can discover devices and modify real-world lighting state on the local network, so accidental invocation may lead to unintended control of multiple devices or exposure of local network device metadata.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
80% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · references/workflow.md (reported line 23)May include surrounding context.

md
Use saved friendly references first. Resolve aliases, rooms, groups, and scenes to
stable device IDs inside the runtime. Refresh stale saved targets before a write. A
new endpoint requires the opaque rebind challenge; do not ask the user to compare raw
IP addresses. If the same ID appears from different senders in one discovery window,
report an identity collision and do not import it.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

This code returns multiple hard-coded Chinese error messages such as at L021, L023, and L028, and the same pattern continues throughout the file. Because the skill provides no user opt-in, fallback, or documented region-specific justification, it violates the language/locale policy for natural-language content.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The file hard-codes the confirmation phrase as "确认恢复场景操作", which requires a specific language for a safety-critical confirmation flow. This is a natural-language locale policy issue because users who do not operate in Chinese are not given a language choice or any documented justification for the restriction.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

This code file includes natural-language error text in Chinese, such as the message at L0124, and the pattern continues throughout the file. Under the policy rule, forcing a specific language without user opt-in is a locale/language policy violation unless the locale restriction is explicitly justified, which is not present here.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

This code embeds user-facing natural-language strings entirely in Chinese for recommended scene names/descriptions and continues that pattern throughout validation and runtime errors. Under the policy, forcing a specific language without user opt-in is a locale/language policy violation unless the constraint is explicitly justified, which is not present in this file.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

This code explicitly uses the "en-US" locale when validating time zones via Intl.DateTimeFormat. Under the policy, hard-coding a specific language or locale without opt-in or clear regional justification is a natural-language policy violation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The formatter for local date/time output is hard-coded to "en-US". This imposes a locale choice on all users and can violate organizational language/locale policy when no opt-in or justification is provided.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

Weekday derivation uses Intl.DateTimeFormat with the locale fixed to "en-US". Even though the output is used programmatically, this still hard-codes a locale preference without opt-in or stated regional need.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The file embeds user-facing names in Chinese (for devices and later-created resources) and later uses a fixed timezone, which together indicate a locale-specific behavior. The policy forbids forcing a specific language or locale without user opt-in, and this script does not expose any choice or justification in the file.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
96% confidence
Finding

This JSON manifest contains user-facing natural-language values such as scene names and provenance fragments exclusively in Chinese. Because the file does not document that the skill is China-specific or offer any language/locale opt-in, it may violate a language/locale policy requiring user choice or explicit justification.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

The nameKey function lowercases names using toLocaleLowerCase("en-US"), which hard-codes a specific locale in behavior that affects matching and identity resolution. This is a natural-language/locale policy concern because users are not offered a locale choice and no region-specific justification is present in the file.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

The code hard-codes Chinese fallback labels ("未命名设备", "未知型号") for public device output. This imposes a specific language on users without any visible opt-in or locale selection, which matches the language/locale policy concern for natural-language content in code.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
97% confidence
Finding

The fallback error message "操作失败。" is always returned when no string error message is available. Because the file provides no user language choice or documented locale constraint, this is a natural-language locale policy issue.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
90% confidence
Finding

The code assigns the default home name to the Chinese string "我的家" when the caller does not provide a home name. This imposes a specific language choice by default rather than offering a locale-neutral value or user-selected language.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

During reset, the code falls back to the same hardcoded Chinese name "我的家" if neither options.homeName nor the current store name is available. This again enforces a specific language choice regardless of user preference or locale.

Content

No source excerpt is available for this finding.