The Agent Skills Directory

[SAFE]: The skill provides purely instructional content for text transformation. It defines rules for analogies, sentence structure, and vocabulary constraints to ensure the AI's output is easy to understand without introducing executable code or shell commands.- [PROMPT_INJECTION]: The activation signals (e.g., 'ELI5', 'dumb it down') and session persistence rules are intended behavior for the skill's functionality and do not attempt to override core agent safety constraints.- [EXTERNAL_DOWNLOADS]: The documentation mentions external GitHub repositories in the 'Sources' section, but these are provided for attribution and documentation purposes. There are no instructions to download or execute remote code.- [DATA_EXFILTRATION]: No mechanisms for reading sensitive local data, environment variables, or credentials were identified.- [SAFE]: Indirect Prompt Injection analysis: The skill ingests user input signals to determine its operating mode. However, it lacks dangerous capabilities such as file writing, shell execution, or network access, and acts only as a formatting layer for assistant replies.
Ingestion points: User conversation signals in SKILL.md.
Boundary markers: Absent.
Capability inventory: Text-based output formatting only.
Sanitization: Absent.

eli5