ClawHub
Back to skill

Security audit

Uber Eats

Security checks for vulnerabilities and agentic risk

Overview

This is a clearly scoped Uber Eats helper that uses approved browser or app sessions and optional local notes, with explicit confirmation required before live orders.

Install this only if you want an agent to help with your real Uber Eats browsing and checkout flow. Review whether you want optional local memory in ~/uber-eats/; keep it to address labels and preferences, and do not store passwords, full payment details, verification codes, or sensitive receipts.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

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 118)May include surrounding context.

md
- Treating the home page as actionable before an address is set -> merchant availability and fees are unreliable.
- Ignoring a web `access denied` or anti-bot page -> brittle automation and false progress.
- Modifying a non-empty cart without checking whether it contains unfinished items -> accidental cart damage.
- Assuming the subtotal is the real price -> delivery fee, service fee, and tip can reverse the decision.
- Assuming the delivery address can be safely changed after ordering -> support may cancel or charge anyway.
- Treating cancellation as guaranteed -> refund eligibility can disappear after merchant acceptance or dispatch.

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 164)May include surrounding context.

md
- Treating the home page as actionable before an address is set -> merchant availability and fees are unreliable.
- Ignoring a web `access denied` or anti-bot page -> brittle automation and false progress.
- Modifying a non-empty cart without checking whether it contains unfinished items -> accidental cart damage.
- Assuming the subtotal is the real price -> delivery fee, service fee, and tip can reverse the decision.
- Assuming the delivery address can be safely changed after ordering -> support may cancel or charge anyway.
- Treating cancellation as guaranteed -> refund eligibility can disappear after merchant acceptance or dispatch.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
82% confidence
Finding

The skill advises storing reusable session reliability information such as working browser/locale combinations and profile states that trigger blocking, but it does not include any privacy boundary or data-minimization guidance. In a consumer ordering context, this can lead agents or downstream systems to persist user-linked behavioral or device/session metadata longer than necessary, creating avoidable privacy and profiling risk.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill explicitly directs the agent to record activation rules, browser-profile reuse preferences, ordering boundaries, address labels, neighborhoods, merchant preferences, and related behavioral data in persistent memory without clearly requiring informed user consent at the time of collection. This creates a privacy and data-minimization risk because location-linked ordering habits can reveal sensitive lifestyle patterns and may be retained longer than the user expects.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.