The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it processes untrusted technical data retrieved from external sources.\n
Ingestion points: Technical research data is gathered through patent/paper searches and web research tools like Exa, Tavily, Brave, and Jina, as specified in SKILL.md and references/source-routing.md.\n
Boundary markers: The instructions lack requirements for using delimiters or boundary markers to separate external data from agent instructions.\n
Capability inventory: The skill performs local file system writes for multiple research artifacts and utilizes document processing tools for exports, as noted in SKILL.md and references/deliverables.md.\n
Sanitization: There are no explicit instructions to sanitize or validate content retrieved from external sources before integration into the report workflow.\n- [COMMAND_EXECUTION]: The skill suggests using established document processing utilities for generating reports.\n
Evidence: references/deliverables.md mentions the optional use of pandoc, python-docx, and reportlab to render Markdown into Word or PDF documents.\n- [EXTERNAL_DOWNLOADS]: The skill utilizes well-known research and data retrieval services.\n
Evidence: SKILL.md and references/source-routing.md document the use of services like Exa, Tavily, Brave, and Jina for technical discovery and deep reading of web content.

tech-route-comparison