The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to interact with a user's codebase and test suite, creating a surface where malicious instructions embedded in project files could influence the agent's behavior.
Ingestion points: The agent uses Read, Grep, and Glob tools to ingest source code, test files, and configuration (e.g., test/, config/, .env).
Boundary markers: No specific boundary markers or instructions to ignore embedded commands in project data are defined in the skill instructions.
Capability inventory: The skill allows file modification (Write, Edit) and shell command execution (Bash) for running Rails tests.
Sanitization: There is no evidence of sanitization or filtering of content read from the project files before it is processed or used to generate new code/tests.
[EXTERNAL_DOWNLOADS]: The skill references standard dependencies and services for a Rails environment. These references are documented neutrally as they target well-known or trusted services.
The skill suggests the use of established Ruby gems such as simplecov, cuprite, selenium-webdriver, capybara, vcr, and webmock in references/ci.md, references/fixtures.md, and references/system-tests.md.
CI configuration in references/ci.md includes standard GitHub Actions (actions/checkout, ruby/setup-ruby, actions/upload-artifact) and Docker images for well-known services (PostgreSQL, Redis).
[COMMAND_EXECUTION]: The skill enables shell execution via Bash for standard Rails testing tasks. The allowed-tools in SKILL.md restrict this to specific commands: bin/rails test:*, bin/rails test, and bundle exec rails test:*. These are appropriate for the skill's stated purpose.

testing