The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute a local validation script (scripts/validate_references.py) to verify the structure and citations of generated slide blueprints. This script is a benign linter that uses standard Python libraries to perform regex-based text analysis and domain whitelisting.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest external user data such as technical briefings, project requirements, and code snippets. It mitigates potential prompt injection risks by enforcing a rigid Markdown structure defined in references/blueprint-template.md and requiring a validation step that checks all citations against a specific list of approved technical domains.
[EXTERNAL_DOWNLOADS]: The skill performs web searches and retrieves technical content from a predefined catalog of reputable sources. These sources include trusted organizations and well-known technology services such as Anthropic, OpenAI, IBM, Oracle, and Hugging Face. These activities are transparently documented and aligned with the skill's primary purpose of technical research.

prepare-technical-slides