The Agent Skills Directory

[SAFE]: The skill provides instructional content for text summarization without executable code or malicious patterns.
[EXTERNAL_DOWNLOADS]: The skill references grounding material and technical specifications from well-known academic sources (Purdue OWL), research repositories (arXiv), and the author's own GitHub repositories. These references are documented as established technical and instructional sources for the summarization task.
[INDIRECT_PROMPT_INJECTION]: The skill inherently processes external text for condensation, which constitutes a standard data-processing surface. The instructions mitigate risk by requiring evidence preservation and faithfulness to the source material.
Ingestion points: External prose provided for summarization as described in SKILL.md.
Boundary markers: Instructions include verification steps to ensure summaries do not invent claims, though specific technical delimiters are not mandated.
Capability inventory: The frontmatter restricts the agent to Read, Grep, and Bash tools.
Sanitization: The skill focuses on semantic preservation; input sanitization is expected to be handled by the agent's core safety protocols.

summarization