The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes legitimate CLI commands such as x writer and the @zh alias to process text. These commands are consistent with the documentation for the x-cmd environment.
[DATA_EXFILTRATION]: The skill provides patterns for reading local files (e.g., article.txt, draft.md) using standard Unix utilities like cat or the -f flag. This file access is intentional and required for the skill's primary function of text processing.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted text data from external files or standard input.
Ingestion points: Local files accessed via the -f flag or piped input (stdin) as described in SKILL.md and SKILL.zh.md.
Boundary markers: None explicitly defined in the provided instructions; the skill relies on the underlying AI processing module.
Capability inventory: The skill uses the x writer tool to send text to an AI model for processing.
Sanitization: No specific sanitization or filtering logic is documented within the skill instructions.

writer