The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses local tools like grep and glob to scan the codebase for specific Rails patterns. It also instructs the agent to run the application's test suite (bundle exec rspec or bundle exec rails test) to establish a baseline before making changes. These actions are standard for development assistance tools and are conducted entirely within the local environment.\n- [EXTERNAL_DOWNLOADS]: The workflow queries the RubyGems API (rubygems.org) to identify the latest patch versions for Rails releases. This is a well-known and trusted service in the Ruby community. Additionally, the skill links to supplementary skills hosted under the author's official GitHub organization (OmbuLabs), which are recognized as trusted vendor resources.\n- [DATA_EXFILTRATION]: The skill accesses project configuration files such as Gemfile.lock and various files under config/ to perform its analysis. This data access is essential for the upgrade process and is handled locally by the agent to generate reports for the user.\n- [PROMPT_INJECTION]: As the skill processes a wide range of project files, it possesses an attack surface for indirect prompt injection. However, it mitigates this by using structured reporting templates and fenced code blocks for any code snippets included in its outputs, which are intended for review by the developer. Ingestion points: Full project codebase and configuration files. Boundary markers: Markdown code blocks and structured reporting templates. Capability inventory: Tool execution (Grep, Glob, Read), shell execution for tests, and generation of bash scripts. Sanitization: Pattern-based detection focusing on known Rails breaking changes.

rails-upgrade