ClawHub
Back to skill

Security audit

Citation Fixer

Security checks for vulnerabilities and agentic risk

Overview

This skill is a straightforward citation-audit helper that tells the agent to review notes, draft fixes, and require user approval before editing.

Before installing, understand that this skill is meant to inspect knowledge base notes for citation issues. Use it on note folders you are comfortable having reviewed, and keep batch edits behind explicit approval as the skill already directs.

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 63)May include surrounding context.

md
## Anti-Patterns

- Fabricating a citation for an uncited claim — flag it as missing, don't invent
- Overwriting a claim citation without checking the source still exists
- Being too aggressive with ad-hoc conversational facts that don't need formal provenance (preference, opinion, speculation clearly marked as such)
- Running the full audit on the entire knowledge base without prioritizing — start with high-traffic pages
- Editing notes directly without the user's approval on the proposed changes

Static analysis

No suspicious patterns detected.