Using LLMs to build tools for iOS location data analysis
Researchers used an LLM to build a tool that decoded iOS location data, overcoming LLM analysis limitations.
- Use LLMs to write specific analysis tools (e.g., parsers, extractors). - Validate AI-generated tool outputs rigorously. - Combine AI tool generation with manual verification for complex data analysis.
Researchers needed to analyze data formats used by iOS Location Services, which are complex and difficult to reverse engineer.
They used Claude Opus to write a Python tool called Extractor that could reliably recover protobuf schemas from Apple binaries. This tool was then used to decode and analyze network traffic from a jailbroken iPhone.
The Extractor tool successfully decoded iOS location services network traffic, demonstrating the utility of AI in creating analysis tools even when direct AI analysis is unreliable.
Researchers used Claude to build a Python tool, Extractor, that successfully recovered and analyzed protobuf schemas from iOS location services data. While direct LLM analysis of the data proved unreliable, the AI-generated tool enabled detailed insights.
By focusing the LLM on code generation for a specific, verifiable task (schema extraction) rather than direct data interpretation, the researchers could create a reliable analysis tool. Validation steps ensured the tool's accuracy.
Do not use LLMs for direct, unverified analysis of sensitive or complex data formats.
- Decoding What iOS Location Services Send to Apple: An LLM Is an Unreliable Analyst but a Useful Toolmaker ↗arXiv · arXiv cs.AI · Oct 15, 2026 · paraphrased
First seen on Subvolts Oct 15, 2026. Updated Oct 19, 2026. Spotted a problem? Tell us.
- Decoding What iOS Location Services Send to Apple: An LLM Is an Unreliable Analyst but a Useful ToolmakerAnthropic · arXiv cs.AI · Oct 15, 2026
The Playbook: reusable how-tos, prompts, configs and tips pulled out of the news.
- ToolExtractor tool to recover protobuf schemas from Apple binariesPython tool to reliably recover protobuf schemas from Apple binaries for analysis. · 30 min
- HeuristicLLMs are unreliable for direct privacy analysis but good toolmakersUse LLMs to build analysis tools, not for direct forensic or privacy analysis. · 15 min
- HeuristicAuxiliary environment signals improve LLM authorization accuracyProviding LLMs with auxiliary environment signals can substantially improve their accuracy in authorization decisions. · 30 min
- ToolMoMHa for multi-objective LLM harness optimizationOptimizes LLM harnesses for accuracy, safety, and token cost using a joint-reward proposer. · 120 min
- TemplatePrompt Library for Life Sciences TasksA collection of pre-written prompts designed for optimal results on common life sciences tasks. · 2 min
- ChecklistEvaluating LLM N-day exploit generationSteps to assess an LLM's ability to turn security patches into proof-of-concept crashes. · 30 min
Questions people ask
- Which AI tools does this work with?
- Any current model. It was seen with Claude in the sources.
- Where did this come from?
- arXiv (arXiv cs.AI). Subvolts collected, normalized and credited it; the link above opens the original.