to-ios-probe-script
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to process external application source code to resolve UI identifiers and Objective-C classes, creating an attack surface for indirect prompt injection.\n
- Ingestion points: Application source code and identifier definitions used to populate Driver and Evidence generation patterns as specified in
SKILL.mdandreferences/driver-generation.md.\n - Boundary markers: The skill relies on specific code generation templates and instructions to the AI but lacks technical delimiters or explicit instructions to ignore instructions found within the processed source code.\n
- Capability inventory: The skill generates code capable of hooking runtime methods (
Override.run), extracting sensitive application state (withStateEvidence), and performing direct memory manipulation via Frida Gum APIs.\n - Sanitization: The instructions recommend sanitizing NSLog TAG evidence in
references/critical-chain-evidence.mdbut do not provide a mechanism to sanitize the untrusted application code that triggers the generation process.
Audit Metadata