The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires and references external tools and documentation for its core functionality.
Suggests the installation of a third-party CLI tool (belt-sh/cli) via npx.
References external installation guidelines hosted on GitHub: https://raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to chain multiple AI models together (e.g., research -> script -> video), which creates a vulnerability surface where a malicious payload from an upstream source could influence downstream actions.
Ingestion points: The pipeline ingests data from external sources such as web search results via tavily/search-assistant and generated scripts from LLMs like openrouter/claude-sonnet-45.
Boundary markers: There are no explicit delimiters or instructions provided in the pipeline commands to prevent the agent from interpreting instructions that might be embedded in the output of these external models.
Capability inventory: The skill has the capability to execute shell commands via the belt CLI, write to local files (e.g., > script.json), and perform network operations through the inference platform.
Sanitization: No sanitization or validation logic is present to filter content generated by upstream models before it is passed as input to subsequent video, image, or audio generation tools.

ai-content-pipeline