The Agent Skills Directory

[SAFE]: The skill is purely instructional and follows established software engineering best practices for code maintainability. No signs of obfuscation, hardcoded credentials, or exfiltration were detected.
[COMMAND_EXECUTION]: The instructions utilize standard local development tools to facilitate code analysis:
Ingestion points: Analyzes local source code using ripgrep (rg) to identify name occurrences.
Tools: Employs eslint and ruff for linting, and local scripts tools/cc-scan.py and tools/arch-scan.py for detecting magic numbers and negative conditionals.
Context: All operations are performed within the local repository environment as part of a refactoring workflow.
[INDIRECT_PROMPT_INJECTION]: The skill involves reading and refactoring external source code, which constitutes an ingestion point for untrusted data.
Ingestion points: Code snippets or files provided by the user for refactoring.
Capability inventory: Includes local file search (rg) and static analysis (linters); no network exfiltration or persistent writing capabilities are triggered by the skill itself.
Boundary markers: The skill uses a defined "Definition of Done" and safe rename procedures to constrain the agent's behavior.
Sanitization: Not explicitly mentioned, but the focus is on renaming via IDE/LSP tools which inherently limits execution risk.

clean-code-naming