The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes user-supplied Chinese patent applications and invention disclosures. This creates an inherent surface for indirect prompt injection if the source documents contain adversarial instructions intended to influence the agent's output.
Ingestion points: Authoritative Chinese source materials, including claims, descriptions, and disclosures provided by the user (refer to SKILL.md).
Boundary markers: The skill does not explicitly define delimiter-based boundary markers (e.g., XML tags) for the untrusted input text.
Capability inventory: The skill uses MCP tools (patent_briefing, advanced_patent_search) and performs file system operations to organize project directories.
Sanitization: No explicit sanitization or input filtering is mentioned, though the skill mandates strict faithfulness to the source text, which serves as a task-specific constraint.
[EXTERNAL_DOWNLOADS]: The skill references external Model Context Protocol (MCP) servers located on the author's official platform.
Evidence: Documentation in README.md and SKILL.md points to patent_briefing and advanced_patent_search servers at https://open.patsnap.com.
Context: These resources belong to the skill author (Patsnap) and provide legitimate, intended functionality for bibliographic data retrieval and patent search. The skill correctly instructs users to manage API keys securely using the client's credential mechanisms rather than hard-coding them.

translate-patents-for-foreign-filing-ip