The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides links to download application binaries and clone source code from the project's GitHub repositories (e.g., github.com/mascanho/RustySEO). These are standard distribution methods for the described software.
[COMMAND_EXECUTION]: Development and installation instructions include standard commands such as npm install, npm run tauri, and cargo build. These actions are necessary for building and running the tool from source.
[INDIRECT_PROMPT_INJECTION]: The toolkit ingests data from external websites and server logs which is then processed using AI models for insights. This represents a potential surface for indirect prompt injection, which is a common risk for tools that analyze external content.
Ingestion points: shallow_crawl and deep_crawl functions in SKILL.md which fetch external web content.
Boundary markers: No explicit delimiters or instruction guards are documented in the provided code snippets.
Capability inventory: The skill uses invoke('ai_chat', ...) and invoke('ai_generate_topics', ...) to interact with LLM providers using crawled data.
Sanitization: No specific sanitization or validation steps for ingested data are described in the instructions.
[PRIVILEGE_ESCALATION]: Troubleshooting documentation advises users on how to bypass operating system security warnings for unsigned software (e.g., macOS 'Open Anyway'). While common in community-developed tools, this practice involves lowering system security protections.
[DATA_EXPOSURE]: The skill is designed to access local server logs (e.g., /var/log/nginx/access.log) and store data in a local SQLite database (~/.rustyseo/data.db). These operations are consistent with the tool's stated purpose of SEO log analysis and history tracking.

rustyseo-toolkit