ClawHub
Back to skill

Security audit

Goal Tracker

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed Markdown-only goal-tracking skill that edits local goal files when used and does not show hidden code, exfiltration, or deceptive behavior.

Before installing, decide whether you want the agent to update goal files automatically during check-ins. Use explicit prompts such as "Weekly goal check-in" and review changes to goals/GOALS.md, retrospectives, and archives before relying on them. Only add the heartbeat or cron examples if you want scheduled reminders sent to your configured chat.

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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (2)

Vague Triggers

Medium
Confidence
93% confidence
Finding
The trigger phrase "How are my goals looking?" is broad enough to overlap with ordinary conversation, which can cause the skill to activate when the user is merely asking a general question rather than intentionally invoking the workflow. In a skill that reads and updates local goal files, unintended activation can lead to unexpected file modifications, noisy check-ins, or disclosure of goal status in the wrong context.

Vague Triggers

Medium
Confidence
95% confidence
Finding
The automatic trigger via heartbeat or cron is underspecified and lacks scope, timing, and safety constraints, so the skill could run unattended in contexts where the user did not expect it. Because the protocol includes reading and updating `goals/GOALS.md`, ambiguous automation increases the risk of repeated unsolicited actions, accidental state changes, and cross-context interference with other workflows.

Static analysis

No suspicious patterns detected.