ClawHub
Back to skill

Security audit

Research Brief

Security checks for vulnerabilities and agentic risk

Overview

This skill provides a clearly scoped, read-only research workflow without requesting unrelated access or actions.

Set a clear research question and search limit. Consider what information external search queries may disclose when researching private topics.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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)

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 92)May include surrounding context.

md
- Searching before defining freshness and the stopping rule, then collecting sources without a clear finish.
- Listing links after drafting without recording which source supports each factual claim.
- Treating a recent page as applicable without checking its product version, policy period, dataset release, or standard revision.
- Hiding contradictory sources inside a blended conclusion.
- Marking a support pass independent when the same seat performed both drafting and checking.
- Turning a read-only brief into an unrequested repository edit, publication, message, or memory entry.

Static analysis

No suspicious patterns detected.