ClawHub
Back to skill

Security audit

Vendor Summit Report

Security checks for vulnerabilities and agentic risk

Overview

The skill coherently creates conference reports, with disclosed optional sharing that warrants checking the publication destination.

Use attachment delivery unless you want distribution. Before optional sharing, review the existing push script and confirm the WeCom group, Pages repository, URL, and report contents. Pages may be public; check [GitHub's visibility controls](https://docs.github.com/en/pages/getting-started-with-github-pages/changing-the-visibility-of-your-github-pages-site).

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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • 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
Findings (1)

The publishing workflow uploads the report to GitHub Pages without warning that its contents may become publicly accessible.

Medium
Category
Output Handling
Confidence
92% confidence
Finding

The optional publishing workflow directs the agent to deploy report HTML to GitHub Pages without explicitly warning that it may become publicly accessible or requiring approval for that public destination. Although the skill targets external readers and primarily uses public conference information, generated reports could include unpublished analysis or user-supplied confidential material; the explicit confirmation requirement covers the WeCom group, not clearly the Pages deployment.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.