Back to skill

Security audit

Personal Knowledge Base

Security checks for vulnerabilities and agentic risk

Overview

This is a plain Markdown knowledge-base workflow with disclosed local note storage and minor data-handling cautions, but no hidden code or exfiltration behavior.

Install only if you want the agent to save content you provide into a persistent local ~/kb/ Markdown folder. Consider asking it to confirm before creating the folder, fetching links, or removing processed inbox items, and keep backups if the notes are important.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (4)

Vague Triggers

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The line says "User sends anything" followed by very broad examples, indicating the skill may activate on almost any message, including common everyday speech. The file does not provide clear boundaries or negative examples to distinguish intentional capture from normal conversation.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

The instruction to delete items from inbox.md once processed authorizes destructive modification of user data without an explicit safeguard, backup, or confirmation step. If processing is incorrect or partial, valuable user content may be permanently lost, making this especially dangerous in a knowledge-base tool whose primary purpose is preservation of information.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The description says the skill helps organize "whatever they send," which is an overly broad trigger scope for a markdown skill description. This could overlap with ordinary conversation and make it unclear when the skill should activate versus when it should not.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The skill instructs creation of a persistent workspace in the user's home directory (~/kb/) without explicit consent, warning, or discussion of data placement. Writing files to a home-directory path can create privacy, data-governance, and surprise-persistence risks, especially if the user did not intend local storage or if sensitive content is captured automatically.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.