The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary workflow involves processing external literature and research data retrieved via the research-lookup tool to create manuscript outlines and prose. This creates a vulnerability to indirect prompt injection if the source data contains malicious instructions.\n
Ingestion points: External papers and datasets processed through the research-lookup dependency (SKILL.md).\n
Boundary markers: The instructions lack delimiters or safety warnings to ignore embedded commands within the research material.\n
Capability inventory: The skill utilizes Read, Write, Edit, and Bash tools, allowing for file system modifications and shell command execution.\n
Sanitization: There is no mention of filtering or sanitizing the data obtained from external research sources.\n- [COMMAND_EXECUTION]: The skill instructs the agent to run a shell command to execute a Python script for generating scientific schematics.\n
Evidence: python scripts/generate_schematic.py "your diagram description" -o figures/output.png in SKILL.md.\n
Concern: The script scripts/generate_schematic.py is missing from the provided skill files, preventing a full security review of the code being executed in the environment.

scientific-writing