The Agent Skills Directory

[Prompt Injection] (SAFE): No instructions were found that attempt to override system prompts or bypass AI safety constraints.\n- [Indirect Prompt Injection] (LOW): The skill provides example code for file operations that ingests untrusted strings, creating a potential path traversal surface.\n
Ingestion points: The path and content string arguments in the Rust read_file and write_file commands.\n
Boundary markers: Absent in code snippets; however, the text provides context for their use.\n
Capability inventory: Demonstrates capabilities for reading and writing files on the host filesystem using std::fs.\n
Sanitization: Absent in snippets, though the text explicitly instructs: 'Sanitize file paths and user inputs'.\n- [Data Exposure & Exfiltration] (SAFE): No hardcoded secrets, sensitive file paths, or network exfiltration logic was identified.\n- [Obfuscation] (SAFE): No use of encoding, hidden characters, or homoglyphs was detected.\n- [External Downloads] (SAFE): The mentioned dependencies (Tauri, React, Tailwind) are standard and reputable libraries for the stated development purpose.

tauri-development