The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to run local Python scripts (scripts/generate_interdisciplinary.py and scripts/generate_interdisciplinary_pptx.py) for document creation. This is a standard and safe use of command execution within the skill's defined scope.
[DATA_EXFILTRATION]: The skill interacts with the file system via established platform paths (e.g., /mnt/user-data/outputs/). Any network-based operations, such as uploading to Google Drive, are conducted through official platform MCP connectors and require explicit user consent.
[PROMPT_INJECTION]: Analysis identified a surface for indirect prompt injection in SKILL.md where user-provided curriculum details are used as shell command arguments. While this represents a common architectural pattern for such skills, it is documented here for completeness.
Ingestion points: User-provided strings for core questions and subjects in SKILL.md Step 1.
Boundary markers: Absent; the skill does not use specific delimiters or instructions to sanitize user input before interpolation.
Capability inventory: Bash (script execution), Write (outputting files to /mnt/user-data/), and Read (accessing local references).
Sanitization: No validation or escaping of user input is present in the bash command templates.

tw-edu-interdisciplinary