The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on a custom rota CLI to perform core functions, such as querying knowledge bases (rota knowledge), managing issues (rota tracker), and performing refactoring tasks (rota refactor). This toolset is central to the skill's operation within the project environment.
[INDIRECT_PROMPT_INJECTION]: The skill has a broad data ingestion surface, as it reads codebase files and metadata from issue trackers to identify architectural friction. Because this data can originate from untrusted sources (such as third-party code or comments), it creates a vulnerability where embedded instructions could influence the orchestrator agent's ranking of findings or the worker agent's code modifications.
[DYNAMIC_EXECUTION]: The verification process in fix-path.md instructs the agent to execute shell commands listed in the test.full array within .rota/config.json. These commands are executed verbatim to validate refactorings, which could lead to arbitrary command execution if the project's configuration file is maliciously altered.

rota-refactor