The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user content, including SME notes, lesson outlines, and existing documents, which are processed to generate video chapters and narrations. This creates a surface where malicious instructions hidden in the data could potentially influence agent behavior.
Ingestion points: Lesson outlines and SME notes mentioned in the 'Inputs' section of SKILL.md.
Boundary markers: No specific delimiters or 'ignore' instructions are provided to separate user content from system instructions.
Capability inventory: The skill uses Clueso MCP tools to create and edit video projects, export files, and generate review links.
Sanitization: No explicit sanitization or validation of the input material is described in the workflow.
[COMMAND_EXECUTION]: The skill instructions allow the agent to offer to run shell commands to configure the Model Context Protocol (MCP) server for environments like Claude Code.
Evidence: The skill specifies the command claude mcp add --transport http Clueso https://connect.clueso.io/mcp to be run with user confirmation.
[EXTERNAL_DOWNLOADS]: The skill references and downloads configuration from external sources to enable its functionality. These resources originate from the vendor's official infrastructure.
Evidence: Fetches MCP configuration from https://connect.clueso.io/mcp and references specific plugin IDs for Codex and ChatGPT.

course-module-video