ClawHub
Back to skill

Security audit

iPhone

Security checks for vulnerabilities and agentic risk

Overview

This is a local, instruction-only iPhone troubleshooting guide with disclosed local memory and user-confirmed device changes.

Before installing, know that this skill may keep local notes about your iPhone model, preferences, recurring problems, and approved actions under ~/iphone/. Use it for guided troubleshooting and optimization, and require confirmation before deletions, resets, profile changes, privacy changes, or automations that affect your device behavior.

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
  • 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
Findings (6)

Vague Triggers

Medium
Confidence
86% confidence
Finding
The 'When to Use' section says to use the skill whenever the user wants an 'iPhone copilot experience' and for broad situations like privacy hardening, connectivity failures, and routine optimization. This is expansive and does not define exclusion conditions or clear limits, which could cause unintended invocation for many ordinary support conversations about iPhones.

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
- Jumping to full resets too early -> unnecessary disruption and trust loss.
- Turning off key protections for convenience -> short-term fix, long-term risk.
- Ignoring user rhythm (work, travel, family) -> optimizations do not stick.
- Ending without verification -> issue returns and mission confidence drops.

## Security & Privacy
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.

Vague Triggers

Medium
Confidence
96% confidence
Finding
The template defines auto-activation behavior, default operating modes, and exceptions for asking permission, but leaves all of them blank and unconstrained. In a skill that provides operator-style guidance for privacy, connectivity, and automation on a personal iPhone, this ambiguity can lead downstream implementations to infer broad autonomy and activate or proceed without clear user consent boundaries.

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
- Verification method:

## Risk Boundaries
- Never do without consent:
- Privacy sensitivity level:

## Notes
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.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill explicitly instructs the agent to persist internal notes in `~/iphone/memory.md` without any requirement to disclose this to the user or obtain consent. Because the notes include behavioral preferences, recurring incidents, and approval boundaries, the file can accumulate sensitive operational and privacy-related context that the user may not realize is being stored.

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
## Safety Rules

- Never automate destructive actions without confirmation.
- Keep a manual override path.
- Document what each shortcut changes and why.
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.

Static analysis

No suspicious patterns detected.