The Agent Skills Directory

[Metadata Poisoning] (MEDIUM): There is a significant discrepancy between the documentation and the implementation. SKILL.md defines the skill as a tool for adapting instructional design to cultural pedagogical traditions (Confucian, Ubuntu, etc.). However, the metadata in skill.py (description="learning.pedagogical.traditions for globalization and compliance") and the default parameters (compliance_framework="GDPR") describe a compliance-focused tool. This inconsistency can lead to the agent misinterpreting the skill's capabilities or purpose.
[Indirect Prompt Injection] (LOW): The skill is designed to ingest and transform external content (source_content).
Ingestion points: The source_content parameter in skill.py and the --content CLI flag.
Boundary markers: None identified. There are no delimiters to separate instructions from untrusted content.
Capability inventory: Data transformation and JSON artifact generation (f"{skill_dir}_output.json").
Sanitization: No validation or sanitization is performed on the input content before processing.
[Implementation Error] (INFO): The execute method in skill.py references an undefined variable skill_dir in the processed result string and the artifacts list. This will trigger a NameError at runtime, rendering the skill unusable in its current state.

learning-pedagogical-traditions