The Agent Skills Directory

[SAFE]: The skill exhibits professional engineering standards, utilizing a modular architecture to separate data models, impact logic, and utility functions, which facilitates transparency and maintainability.- [SAFE]: Input data is strictly validated against a comprehensive JSON Schema (Draft 2020-12) via the validate_injurious.py script. This provides strong sanitization against malformed inputs or unexpected data structures.- [EXTERNAL_DOWNLOADS]: The skill depends on the jsonschema Python library to perform its validation tasks. This is a widely recognized and trusted industry-standard library for JSON structure verification.- [INDIRECT_PROMPT_INJECTION]: The skill processes external JSON files, which represents an untrusted data ingestion surface. 1. Ingestion points: Data is loaded via load_from_json in injurious_affection_calculator.py and load_input in validate_injurious.py. 2. Boundary markers: Not applicable, as the data is processed by local scripts rather than being interpolated into LLM prompts. 3. Capability inventory: The skill performs file read/write operations but lacks network access, subprocess execution, or dynamic code evaluation (eval/exec). 4. Sanitization: All inputs are constrained by the JSON Schema, which limits data types and value ranges.- [SAFE]: The code utilizes defensive programming techniques, such as the safe_divide function in utils/calculations.py, to ensure runtime stability and prevent errors like division by zero.

injurious-affection-assessment