The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest data from external, untrusted sources including web search results and various user-uploaded files (JSON, CSV, documents, images) to formulate video generation prompts.
Ingestion points: SKILL.md (mandatory web search), references/discovery-workflow.md (asset inventory), and references/intelligent-inference.md (uploaded file analysis).
Boundary markers: The prompt output format specified in references/prompt-engineering.md lacks explicit delimiters or instructions to help the downstream agent distinguish between core instructions and data-sourced content.
Capability inventory: The skill utilizes web search tools and file system read capabilities to gather context.
Sanitization: No explicit sanitization or validation logic is provided to filter potential malicious instructions embedded within the external content before it is interpolated into the final prompt.
[EXTERNAL_DOWNLOADS]: The skill includes a script that makes a network connection to fetch metadata from a remote server.
Evidence: scripts/check-updates.py uses urllib.request to query https://registry.npmjs.org/remotion/latest.
Context: This operation fetches version information from the official npm registry, which is a well-known and trusted service. The finding is documented neutrally as a safe operation.
[COMMAND_EXECUTION]: The skill package includes a standalone Python utility intended for maintenance and integrity verification.
Evidence: scripts/check-updates.py is provided to check for Remotion updates and verify that all referenced markdown files exist on the local disk.

remotion-prompt-generator