The Agent Skills Directory

[SAFE]: No malicious behavior detected. The skill is designed for pedagogical purposes, producing student-facing scaffolds and teacher guidance.
[PROMPT_INJECTION]: The prompt structure is safe. It defines a specific persona (SRL expert) and provides structured instructions for content generation without attempting to bypass safety protocols or override system behavior.
[DATA_EXFILTRATION]: The skill does not request sensitive file access, environment variables, or perform any network operations that could lead to data exfiltration.
[COMMAND_EXECUTION]: There are no shell commands, subprocess calls, or scripts included in the skill. The logic is entirely contained within the LLM prompt.
[REMOTE_CODE_EXECUTION]: No external dependencies, package installations, or remote script downloads are present.
[INDIRECT_PROMPT_INJECTION]: While the skill ingests untrusted data via fields like task_description and student_profiles, the risk is minimal as the skill's capabilities are limited to text generation for educational materials and it lacks any dangerous system-level access.

self-regulation-scaffold-generator