The Agent Skills Directory

[SAFE]: The skill establishes a robust workflow for generating technical explanations tailored to different audience levels. It incorporates pedagogical frameworks like Bloom's Taxonomy and the Concrete-Representational-Abstract (CRA) model to ensure educational clarity and structured knowledge transfer.
[PROMPT_INJECTION]: The skill processes external code and architectural descriptions via the /analyze:explain command, creating a surface for potential indirect prompt injection. Ingestion points: The [target] parameter in references/explain.md passed to Read/Grep/Glob and subagent prompts. Boundary markers: The workflow lacks explicit delimiters or escape sequences when interpolating target content into reasoning prompts. Capability inventory: The skill uses Read/Grep/Glob for file access, Task for delegation, and Write for documenting results. Sanitization: No automated filtering or sanitization of input code is performed. However, the workflow includes explicit instructions for human-like validation of accuracy and prohibits revealing sensitive details.

code-explanation