The Agent Skills Directory

[SAFE]: The skill and its associated reference playbook focus exclusively on legitimate software engineering tasks, such as code analysis, refactoring strategy, and the application of design patterns (SOLID). All code samples provided are educational examples for 'Before/After' comparisons.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided code through the $ARGUMENTS placeholder in SKILL.md. While this ingestion of untrusted data is a potential vector for indirect prompt injection, the skill's purpose is to perform static analysis and refactoring on that data, which is a benign and intended functionality.
[COMMAND_EXECUTION]: The instructions encourage validating refactored code with tests and regression checks. This implies the agent may use local testing frameworks (like pytest or npm test), which is standard behavior for a code-editing skill.
[EXTERNAL_DOWNLOADS]: The implementation-playbook.md file contains numerous references to well-known industry tools, GitHub Actions, and static analyzers (e.g., SonarQube, Semgrep, ESLint, Ruff). These are documented as best-practice recommendations for the user's environment and do not trigger any unauthorized downloads or remote script executions by the skill itself.

codebase-cleanup-refactor-clean