The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted property data from JSON files and incorporates it into reports for analysis by the AI agent. This ingestion creates a potential surface for indirect prompt injection if input descriptions contain adversarial instructions.
Ingestion points: modules/validators.py and encumbrance_discount_calculator/modules/validators.py load property and encumbrance data from user-provided files.
Boundary markers: Absent; the data is interpolated into Markdown headers and lists in the generated reports without strict delimiters.
Capability inventory: The skill possesses file-writing capabilities to save analysis reports in Markdown and JSON formats (modules/output_formatters.py).
Sanitization: No explicit sanitization or filtering is performed on descriptive text fields before they are presented to the agent.

title-expert