chaptered-training-video
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill includes instructions to execute the
claude mcp add --transport http Clueso https://connect.clueso.io/mcpcommand. This command installs an external Model Context Protocol (MCP) server from the vendor's domain. While this is a documented setup procedure for the skill's functionality, it involves executing shell commands that modify the agent's execution environment and capabilities. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data by transcribing and analyzing user-provided video recordings to identify topic boundaries. This creates a surface for indirect prompt injection attacks where malicious instructions embedded in a video's audio could attempt to influence the agent's behavior.
- Ingestion points: Audio transcripts extracted from the user's video recordings (referenced in
SKILL.mdunder the 'Find chapter boundaries' section). - Boundary markers: The skill does not define specific delimiters or instructions to treat the transcript content as data rather than instructions.
- Capability inventory: The skill possesses significant capabilities via the
clueso-mcp, including project manipulation, timeline editing, file generation (title cards), and project export. - Sanitization: There is no evidence of sanitization or filtering of the transcript content before it is processed by the LLM for structural analysis.
Audit Metadata