ios-watchos

Warn

Audited by Gen Agent Trust Hub on Mar 7, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references an external Swift Package Manager dependency at 'https://github.com/openclaw/sdk.git'. This repository and the associated 'openclaw' organization are not recognized as trusted or well-known entities.- [EXTERNAL_DOWNLOADS]: The skill configures network operations to an external WebSocket service at 'wss://api.openclaw.com/watch'. Use of unknown third-party endpoints for sensitive application data poses a risk if the service is compromised.- [COMMAND_EXECUTION]: Instructions include the use of 'xcodebuild' commands for CI/CD pipelines. While standard for iOS development, these provide a mechanism for arbitrary command execution if build parameters are influenced by untrusted inputs.- [PROMPT_INJECTION]: The skill implements data ingestion via 'WatchConnectivity' ('WCSession.default.didReceiveMessage') and the 'OpenClaw API' ('didReceiveData').
  • Ingestion points: SKILL.md (Usage Pattern 3, 4, 6, Common Commands/API sections).
  • Boundary markers: No markers or delimiters are specified to distinguish between trusted instructions and untrusted data.
  • Capability inventory: The skill utilizes network access (WebSocket), health data access (HealthKit), and shell command execution ('xcodebuild').
  • Sanitization: There is no evidence of input validation, escaping, or sanitization for data received from external sources before it is processed by the agent or the application logic.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 7, 2026, 05:44 PM