The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources during the research phase, creating a surface for indirect prompt injection. The instructions do not define specific boundary markers or sanitization for this data. * Ingestion points: Research-lookup skill and web_search/read_url tools mentioned in SKILL.md. * Boundary markers: None specified in the instructions. * Capability inventory: code_execute, create_document, and web_search. * Sanitization: Not explicitly implemented; however, the two-stage drafting process provides a structural buffer.
[COMMAND_EXECUTION]: The skill leverages the code_execute tool to run local Python scripts (scripts/generate_schematic.py and scripts/generate_image.py) for scientific figure generation. This execution is limited to intended functional capabilities.

scientific-writing