The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes several shell commands to identify project rules and files to process, including find for rule discovery, git diff for change detection, and grep/rg for pattern analysis and 'dead mock' identification.
[COMMAND_EXECUTION]: The skill runs language-specific linters and test runners such as rubocop, eslint, ruff, black, and rspec. It also attempts to dynamically discover build and lint tasks from project files like Makefile, package.json, and pyproject.toml if the language is not predefined.
[SAFE]: The skill accesses the user's global configuration file located at ~/.claude/CLAUDE.md to gather coding standards. While this is outside the project directory, it is used for its intended purpose of applying consistent coding rules and is a standard path for the platform.
[PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes untrusted data from project files and follows instructions within CLAUDE.md and rules files. Ingestion points: Workflow steps read CLAUDE.md, .claude/rules/*.md, ~/.claude/CLAUDE.md, and source code files. Boundary markers: Absent; the skill is instructed to follow instructions found in these external rule files. Capability inventory: The skill can execute shell commands (grep, rg, linters, rspec) and invoke sub-agents. Sanitization: Absent.

polish-before-commit