The Agent Skills Directory

[PROMPT_INJECTION]: The skill has a surface for indirect prompt injection (Category 8) because it is instructed to read and analyze external files such as source code, PRDs, and technical documentation.
Ingestion points: Reads local code repositories, 기술 문서 (Markdown, Word), and project requirement documents (PRD) provided by the user.
Boundary markers: Absent. The instructions do not explicitly mandate the use of delimiters or 'ignore embedded instructions' prompts when processing these external files.
Capability inventory: The skill primarily reads files and writes technical documentation to the local archive/ directory.
Sanitization: Absent. There is no instruction to filter or sanitize the contents of the files being read before analysis.
[DATA_EXPOSURE]: The skill is designed to analyze sensitive project structures and configuration templates (e.g., .env.example, package.json). While this is necessary for its stated purpose of patent evidence extraction, users should be aware that the agent will access the codebase and technical design documents. No evidence of unauthorized credential harvesting or exfiltration was found.
[EXTERNAL_DOWNLOADS]: The skill references official patent drafting guidelines from the China National Intellectual Property Administration (CNIPA). These are well-known, trusted government sources used for reference purposes.

code2patent