The Agent Skills Directory

[PROMPT_INJECTION]: Potential surface for indirect prompt injection detected in prompt-engineering-research/SKILL.md.
Ingestion points: Several functions (create_research_prompt, research_cot_prompt, screen_paper_relevance) interpolate external data such as paper abstracts and excerpts into prompt strings.
Boundary markers: The templates do not utilize specific delimiters or guardrail instructions to isolate the untrusted input from the system instructions.
Capability inventory: The skill provides logic for generating prompts that an agent would subsequently process to perform analysis or summarization.
Sanitization: There is no evidence of sanitization or escaping of the external content before it is embedded in the prompt.
[EXTERNAL_DOWNLOADS]: Fetches academic paper metadata from the official arXiv repository using the arxiv Python library.
Evidence: Seen in autonomous-agents-papers-guide/SKILL.md, domain-adaptation-papers-guide/SKILL.md, graph-learning-papers-guide/SKILL.md, and ai-agent-papers-guide/SKILL.md.
This behavior is consistent with the primary purpose of the skills which is to facilitate literature review and research tracking.

ai-ml-skills