ClawHub
Back to skill

Security audit

AI Ethics Compass

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill suspicious because scanners report public publishing workflows alongside ethics guidance, while the review is incomplete.

Review the skill’s public-submission workflows and consider whether you are comfortable with the possibility that submitted content may be published. The activation scope is also broad. Treat the verdict as a caution: Skillspector’s review is partial because two local references were missing, although AIG completed without findings.

A.I.G

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

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

The markdown ethics-reference skill also provides workflows that publish user submissions to remote public resources.

Medium
Category
Excessive Agency
Confidence
94% confidence
Finding

The manifest describes consulting an ethics compass as markdown, but L93-L107 provide a POST workflow that creates a public issue, and L111-L120 provide another that turns a submitted dilemma into a public conversation page. These are externally visible write operations, not merely retrieval of ethical guidance, and the latter can publish potentially sensitive dilemma details.

Content

No source excerpt is available for this finding.

Activation is framed around general decision-making rather than an explicit request for ethical guidance.

Low
Category
Trigger Abuse
Confidence
84% confidence
Finding

The skill is offered for any decision that a rule does not settle, including declining a request or correcting a mistake. These situations occur in ordinary agent work, and the description does not distinguish when to invoke this skill from when to handle them without it.

Content

No source excerpt is available for this finding.