The Agent Skills Directory

[SAFE]: The skill implements a structured methodology for intellectual property research, focusing on novelty and non-obviousness assessments.
[SAFE]: The instructions direct the agent to utilize standard, reputable search engines and patent databases (e.g., Google Patents, USPTO, WIPO, arXiv), which are considered trusted sources for this use case.
[SAFE]: No executable code, obfuscated commands, or unauthorized data access patterns were found in the skill definitions.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits a typical surface for indirect prompt injection by processing external data, though the risk is deemed minimal in this search-oriented context.
Ingestion points: The skill ingests user-supplied invention descriptions and technical content retrieved from the web via the WebFetch tool.
Boundary markers: There are no explicit delimiters defined to separate the skill instructions from untrusted external data in the search results.
Capability inventory: The skill is configured with read-oriented capabilities including WebSearch, WebFetch, Read, Glob, and Grep for data retrieval and analysis.
Sanitization: The instructions do not specify any validation or sanitization steps for the data retrieved from external sources before processing.

search-prior-art