The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions in SKILL.md suggest executing shell commands for behavior validation, specifically npm test, npx tsc, and npm run lint. It also describes a multi-agent workflow using CLI tools such as ask-gemini and codex-cli for codebase analysis and verification.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it processes untrusted code provided by users or external repositories.
Ingestion points: Untrusted source code provided for refactoring as described in the instructions of SKILL.md.
Boundary markers: Absent. The instructions do not specify delimiters or warnings to ignore instructions embedded within the code being refactored.
Capability inventory: The skill encourages the use of shell commands (npm test, npx tsc, npm run lint) which could be exploited if the code being refactored contains malicious configuration files (e.g., a poisoned package.json script).
Sanitization: Absent. There is no mention of sanitizing or escaping the input code before the agent processes or executes tests against it.

code-refactoring