The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions in SKILL.md refer to a local script 'file_metrics.py' for analyzing file size; however, the script is not included in the provided source files.
[EXTERNAL_DOWNLOADS]: The skill includes functionality to fetch content from external URLs using standard web retrieval tools as part of its core summarization purpose.
[PROMPT_INJECTION]: The skill processes untrusted external data, creating an Indirect Prompt Injection surface. Ingestion points: local files, external web URLs, and image files. Boundary markers: the skill enforces structured output formats but lacks explicit input delimiters. Capability inventory: the agent has access to file reading, network requests, and teammate orchestration tools. Sanitization: the skill mandates a 'Extract Before Abstracting' fidelity rule, which requires grounding the summary in verbatim quotes from the source, significantly mitigating the risk of following instructions embedded in the content.

summarizer