The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied videos and potentially user-defined HTML for rendering, which constitutes a standard surface for indirect prompt injection.
Ingestion points: The agent accepts video_urls and HTML configuration from users or external sources as specified in SKILL.md.
Boundary markers: No explicit delimiters or boundary markers are instructed for the agent to distinguish between media metadata and operational instructions.
Capability inventory: The skill has access to media manipulation tools including merge_media, scale_video, trim_video, and generation tools like render_html_video and generate_video via its MCP dependency.
Sanitization: There are no specific instructions for the agent to sanitize or validate metadata from imported clips before processing.
[EXTERNAL_DOWNLOADS]: The skill defines a dependency on a remote Model Context Protocol (MCP) server provided by the vendor.
The agent configuration in agents/openai.yaml specifies a tool endpoint at https://app.creativeclaw.co/mcp/chatgpt to fetch and execute media processing capabilities. This represents the intended communication channel between the agent and the service provider.

creativeclaw-add-video-intro-outro