The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains an attack surface for indirect prompt injection when processing external data. \n- Ingestion points: SKILL.md and references/phase-1-discovery.md instruct the agent to retrieve and analyze YouTube transcripts or web-based content based on user-provided URLs. \n- Boundary markers: The skill does not specify the use of delimiters or instructions for the agent to ignore potentially malicious directions embedded within the fetched transcripts. \n- Capability inventory: The skill utilizes network-search and fetching tools to ingest content and performs local file-write operations to the course-creator-output/ directory. \n- Sanitization: There is no requirement for the agent to sanitize or validate the external content before integrating it into the course generation process. \n- [EXTERNAL_DOWNLOADS]: The skill instructions direct the agent to fetch external data (transcripts) from YouTube or other websites using web tools. While functional, this involves processing content from arbitrary third-party sources. \n- [NO_CODE]: This skill consists solely of instructional markdown and reference files, with no executable scripts, binary files, or external package dependencies included in the skill package.

course-creator