The Agent Skills Directory

[PRIVILEGE_ESCALATION]: The skill includes instructions for users to bypass standard operating system security warnings when installing the desktop application binaries. Specifically, it suggests ignoring 'Unknown Developer' warnings on Windows and using the 'Open Anyway' security override on macOS for the provided .msi and .dmg installers.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection due to the way it handles external data.
Ingestion points: Data is ingested from external sources via website crawling (shallow and deep crawls) and by parsing server log files (Nginx and Apache access logs).
Boundary markers: The instructions do not define clear delimiters or boundary markers to separate untrusted external content from the agent's internal instructions.
Capability inventory: The skill has significant capabilities including network access via Tauri's invoke system, local database access (SQLite), and file system operations for exporting reports.
Sanitization: There is no mention of sanitization or filtering of the content retrieved from external URLs before it is passed to the AI chatbot feature for analysis.
[EXTERNAL_DOWNLOADS]: The documentation provides links to download and execute compiled binaries (.msi, .dmg, .AppImage) and source code from third-party GitHub repositories (mascanho/RustySEO) which are not associated with the predefined list of trusted vendors.

rustyseo-toolkit