The Agent Skills Directory

[SAFE]: The skill performs legitimate code analysis and transformation for accessibility improvements using standard React Native APIs and official documentation references.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface that processes user source code, but includes sufficient safeguards to be considered safe.
Ingestion points: React Native and Expo source files (**/*.{tsx,jsx,ts,js}) are scanned during Phase 1 and 2 to identify existing accessibility implementations and missing properties.
Boundary markers: The skill follows a structured 5-phase workflow focused on identifying specific anti-patterns defined in the rules/ directory.
Capability inventory: The skill uses file read tools for discovery, shell commands (grep, glob) for identification, and the Edit tool for applying automated fixes to the source code.
Sanitization: All automated changes involving inferred labels or hints are explicitly marked with [VERIFY] comments, ensuring that a human developer must review and confirm the changes before they are finalized. This mechanism mitigates the risk of the agent unknowingly applying or executing instructions embedded within the processed files.

react-native-accessibility