The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes several Python scripts (scripts/extract_docx.py, scripts/validate_gzh_html.py, scripts/wrap_preview.py, scripts/component_lint.py) designed to convert documents, validate HTML tags, and generate preview files. These scripts use standard Python libraries (zipfile, re, html.parser) for local file manipulation.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest untrusted data from external sources including .docx, .pdf, .txt, and .md files. This represents a potential surface for indirect prompt injection where instructions embedded in those documents could attempt to influence the agent's behavior. However, the skill explicitly includes a 'structure confirmation' step in its workflow, which allows the user to review the interpreted structure before final processing, mitigating the risk of accidental obedience to embedded instructions.

gzh-design