ClawHub
Back to skill

Security audit

gtrk B-roll 检索铺轨

Security checks for vulnerabilities and agentic risk

Overview

The skill’s media search and editing workflow fits its purpose and clearly discloses cloud processing, charges, and user checkpoints.

Consider whether your media is suitable for cloud processing and review estimated charges. Limit indexing to intended media folders, inspect candidate tracks, and understand that forced replacement can permanently remove prior edits.

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
  • 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)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
91% confidence
Finding

The supplied chunk contains only existing-clip editing commands. It does not search, index, describe, download, or lay candidate media using the declared matrix workflow. Although timeline adjustments could support a broader B-roll process, this chunk supplies standalone patch operations without any connection to that process, making its demonstrated purpose materially different from the declared retrieval-and-placement skill.

Content

No source excerpt is available for this finding.

The skill mandates English search descriptions without user opt-in or a documented justification.

Low
Category
Output Handling
Confidence
96% confidence
Finding

The search-query standard explicitly requires English scene descriptions of 5–12 words for ad-hoc searches and dispatch queries. This is a specific natural-language requirement, not a clarity idiom, and the skill neither offers a language choice nor explains why the search service requires English.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.