The Agent Skills Directory

[SAFE]: The skill operates as a standard developer utility for codebase restructuring. It utilizes local tools and does not perform any suspicious network operations or obfuscated actions.
[COMMAND_EXECUTION]: The skill uses common CLI tools such as git, sed, and build/test runners (e.g., tsc, pytest, cargo) to modify source code and verify the integrity of the system after changes. These commands are typical for the stated purpose of code refactoring.
[DATA_EXFILTRATION]: While the skill audits the local codebase structure to identify modules and hotspots, there is no evidence of reading sensitive files (like .ssh or .env) or sending any data to external servers.
[PROMPT_INJECTION]: The skill identifies a potential attack surface (Indirect Prompt Injection) by reading local source files which could contain instructions intended to influence the agent.
Ingestion points: Reads local files and folder structures during the audit phase as specified in SKILL.md.
Boundary markers: None identified in the prompt templates.
Capability inventory: Includes file modification, shell command execution for refactoring tools, and build system verification commands listed in REFERENCE.md.
Sanitization: Not present; the skill is designed for structural code transformations and does not interpret the content of code comments as instructions.

refactor-codebase