android-tombstone-symbolication
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches debug symbols from Microsoft's official symbol server (
msdl.microsoft.com) and runtime packages from NuGet.org (api.nuget.org). These are well-known and trusted services for software development and diagnostics. - [COMMAND_EXECUTION]: The automated PowerShell script invokes
llvm-symbolizerandllvm-readelfto process binary files. Execution is controlled, using specific offsets and library names extracted from the input tombstone via regular expressions. - [INDIRECT_PROMPT_INJECTION]: The skill processes external tombstone files and logcat output, which represents an attack surface for indirect prompt injection.
- Ingestion points: The skill reads user-supplied crash logs or files via the
-TombstoneFileparameter inscripts/Symbolicate-Tombstone.ps1. - Boundary markers: The script uses strict regex patterns (e.g.,
^\s*#(\d+)\s+pc\s+(0x)?([0-9a-fA-F]+)) to isolate specific backtrace data, providing a degree of structural isolation. - Capability inventory: The skill possesses network access (symbol downloads), file system access (symbol caching in temp directories), and the ability to execute external LLVM binaries.
- Sanitization: Input is filtered through regex to extract hex strings and library names; raw text from the tombstone is not directly executed as shell commands or code.
Audit Metadata