Back to skill

Security audit

学习DNA

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill suspicious with medium confidence because scanners report concerns about consent for persistent student profiles and about following the skill's stated limits; part of the artifact was only partially analyzed.

Consider the reported guardian-consent and instruction-scope concerns before installing, especially if the skill may handle younger students' sensitive profiles. The scan is incomplete: one script was only partially interpreted, and a Unicode flag was disputed, so the results do not establish that the artifact is safe or malicious.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 857)May include surrounding context.

md
- `schemas/validate.js` - **开发期**的 schema 校验脚本:本地读取上面两个文件、编译 JSON Schema、检查示例是否合法。

L020–L023 allow junior-high students to authorize a long-term profile without guardian consent, including students below the manifest's approximately 14-year-old threshold.

High
Category
Excessive Agency
Confidence
98% confidence
Finding

The onboarding template permits student-only consent for all junior-high students without checking whether the student falls below the guardian-consent age threshold. An agent following this workflow could create a persistent profile for a younger minor without required guardian approval, making this omission particularly consequential in a system storing sensitive student information.

Content

No source excerpt is available for this finding.

The skill directly verifies student understanding despite the manifest explicitly excluding understanding verification.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The manifest at L003 limits this skill to storage and authorization and explicitly states that it does not verify understanding; understanding-depth results must arrive through handoffs from the Feynman skill. However, L693-L699 require this skill to challenge an ambiguous claim of understanding and conduct a recall-based verification itself, with L838 and L860 additionally authorizing verification questions.

Content

No source excerpt is available for this finding.

Ae4

Medium
Category
analysis-evasion
Confidence
80% confidence
Finding

Suspicious Unicode normalization or mixed-script content

Content

No source excerpt is available for this finding.