Back to skill

Security audit

OpenClaw Indulgence

Security checks for vulnerabilities and agentic risk

Overview

This skill is mostly a light Chinese-language preference guide, but it includes broad instructions to clean files and add or remove skills without clear user control.

Review this skill before installing. It should be treated as a lightweight preference/philosophy guide, but do not let it authorize automatic file deletion or skill installation/removal unless you explicitly approve each proposed change and affected path.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • 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 (1)

T05 · Unauthorized Access and Privilege Escalation

Warning
Location
SKILL.md:31
Finding

Unbounded File and Skill Removal Instructions

Content
View full analysis

Vulnerability Details

File Location: SKILL.md:31-34; duplicated in README.md:21-24
Vulnerability Type: T05: Unauthorized Access and Privilege Escalation
Risk Level: Medium

Relevant Skill instructions, translated into English:

text
### Organize the Environment
- Clean unused files and streamline workflows
- Check skill matching, remove unused ones, and add needed ones
- Make the entire system clean and organized

Technical Analysis

The Skill instructs an agent to delete files and modify installed Skills without limiting these operations to a user-approved directory or explicitly identified resources. It does not require confirmation, preview proposed changes, protect sensitive paths, define how the agent should determine that an item is unused, or provide rollback and recovery controls.

These mutation instructions are not necessary for the Skill's stated purpose of documenting preferences for relaxation and enjoyable interaction. An agent could interpret the broadly worded cleanup guidance as authorization to operate on project files, configuration, or installed capabilities outside the minimum scope needed for the user's current request.

No executable code, permission-escalation implementation, persistence mechanism, or malicious payload was identified. The risk arises from unsafe agent instructions that could cause unauthorized resource modification if an agent acts on them.

Attack Path

  1. The Skill is loaded into an agent's active context.
  2. The agent interprets environmental organization as an actionable instruction.
  3. Without requesting explicit approval, the agent evaluates files or installed Skills and decides that some are unused.
  4. The agent invokes available filesystem or Skill-management tools to delete files, remove Skills, or add replacement capabilities.
  5. Required data or functionality is lost or altered, potentially affecting later tasks and sessions.
  6. Beca ...[truncated 823 chars]
Remediation
View remediation

Remediation Suggestions

  1. Replace actionable deletion language with read-only organizational recommendations.
  2. Require explicit user approval before every file deletion, Skill removal, Skill installation, or configuration change.
  3. Restrict operations to paths and resources specifically selected by the user.
  4. Present a complete preview of proposed changes, including affected paths and the reason each item was classified as unused.
  5. Define protected locations that must never be modified, including system directories, credentials, configuration, version-control metadata, and unrelated projects.
  6. Prefer recoverable deletion, backups, or version-controlled changes over permanent removal.
  7. Verify that no active workflow or dependency references a file or Skill before proposing its removal.
  8. Require separate confirmation for adding third-party Skills and validate their source and integrity.
  9. Apply the same corrected wording to the duplicated instructions in README.md.
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (3)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The entire skill description is presented only in Chinese, with no indication that other languages are supported or that the language choice is optional. Per the policy, forcing a specific language or locale without user opt-in 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
95% confidence
Finding

The manifest description and all user-facing markdown content are written exclusively in Chinese, with no indication that users may choose another language or that the skill is intended only for a Chinese-speaking context. The policy requires flagging language or locale constraints when they are imposed without user opt-in or clear justification.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The manifest description is written entirely in Chinese and provides no indication that language choice is optional or that the skill is intended only for a Chinese-speaking or region-specific context. This creates a natural-language locale policy concern because the skill implicitly assumes a fixed language without user opt-in.

Content

No source excerpt is available for this finding.