Back to skill

Security audit

mobile-testing

Security checks for vulnerabilities and agentic risk

Overview

ClawScan labels this skill benign because the reported findings appear consistent with mobile testing guidance, though the overall confidence is medium and scanning was incomplete.

The verdict is benign, not risk-free. Before installing, consider that the scan was partial: only 75% of components were scanned, obfuscated instruction text was not fully evaluated, and some references could not be resolved. If using the test-isolation example, make sure the target app’s local data can be deleted.

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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Memory PoisoningPersistent Context Injection, Context Window Stuffing, Memory Manipulation
  • 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 (2)

Memory Manipulation

High
Category
Memory Poisoning
Confidence
80% confidence
Finding

Skill manipulates agent memory, state, or stored context. Memory corruption can alter personality, override safety rules, or cause unpredictable behavior.

Content

Scanner excerpt · references/ui-tests.md (reported line 140)May include surrounding context.

md
- `waitForExistence(timeout:)` and, in current SDKs, `waitForNonExistence(timeout:)` for disappearing
  spinners. No `sleep`.
- System permission alerts: prefer avoiding them with fakes. When the alert is the subject, handle it
  with `addUIInterruptionMonitor(withDescription:handler:)`, or reset state first with
  `app.resetAuthorizationStatus(for:)`. On a simulator, `xcrun simctl privacy <device> grant location
  <bundle id>` pre-grants supported services (Bluetooth is not one of them; the simulator has no BLE radio,
  so use a fake transport).

The test-isolation configuration enables clearing app data without warning that stored app state will be deleted.

Medium
Category
Output Handling
Confidence
90% confidence
Finding

The example sets clearPackageData to true, causing Android Test Orchestrator to clear the app's data between tests. The accompanying explanation mentions restarts and performance overhead but does not disclose deletion of local app data or restrict this configuration to disposable test environments.

Content

No source excerpt is available for this finding.