The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it incorporates untrusted data into its core instructions. \n- Ingestion points: The {{refactor_description}} variable in SKILL.md is directly embedded in the prompt. \n- Boundary markers: The skill lacks delimiters (e.g., XML tags or triple quotes) to isolate the user input from the agent's instructions. \n- Capability inventory: The skill requires searching and analyzing the codebase, which typically involves file-read permissions. \n- Sanitization: No sanitization or filtering is applied to the input variable. \n- Remediation: Wrap the {{refactor_description}} variable in unique delimiters and instruct the agent to ignore any commands contained within it. \n- [NO_CODE]: This skill contains only a markdown instruction file and does not include any Python, JavaScript, or other executable scripts.

refactor-plan