The Agent Skills Directory

[COMMAND_EXECUTION]: The skill requests Bash permissions in SKILL.md to support its principles of ensuring test coverage and committing changes. This provides an execution environment that could be exploited if the agent is misled.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. Ingestion points: The skill ingests and processes untrusted source code provided by the user for refactoring, as indicated in the activation triggers in SKILL.md. Boundary markers: The instructions do not define clear delimiters or instructions to ignore potential commands embedded within the comments of the provided code. Capability inventory: The skill is granted Read, Write, Edit, and Bash tool access. Sanitization: There is no evidence of sanitization or filtering of the input code to prevent the execution of instructions contained within the data.

refactoring-specialist