The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands to inspect the local development environment. It runs version checks (e.g., ruby -v, bundle exec rails -v), git status inquiries, and searches project files (e.g., grep on Gemfile and CI/CD configuration files) to identify current software versions and necessary infrastructure updates.
[DYNAMIC_EXECUTION]: The orchestrator triggers the execution of the project's own verification suites, including RSpec tests and RuboCop linting, to ensure code quality and functionality throughout the upgrade phases. It also utilizes a custom rules engine (rules-engine.md) to apply project-specific upgrade logic.
[INDIRECT_PROMPT_INJECTION]: The skill's logic is influenced by the contents of project files and tool outputs which are treated as untrusted data sources.
Ingestion points: Reads content from Gemfile, Gemfile.lock, .ruby-version, and a local configuration file .ruby-upgrade-toolkit/rules.yml. It also processes raw output from RSpec and RuboCop runs.
Boundary markers: None explicitly specified for the ingestion of project file content or test outputs.
Capability inventory: Includes file reading/writing (Read, Edit, TodoWrite), directory traversal (Glob), and shell command execution (Bash).
Sanitization: The skill performs specific parsing (e.g., regex matching on version strings) to validate data from project files before using it in the upgrade logic.

Upgrade Orchestrator