ai-rag-pipeline
Pass
Audited by Gen Agent Trust Hub on Apr 14, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill directs users to an external installation script for the infsh CLI tool located at raw.githubusercontent.com/inference-sh/skills. This is a standard setup procedure for the platform described.
- [COMMAND_EXECUTION]: The skill extensively uses the Bash tool restricted to the infsh command to perform web searches, content extraction, and LLM orchestration.
- [PROMPT_INJECTION]: The skill facilitates the creation of RAG pipelines that process untrusted data from the web, which is a known vector for indirect prompt injection attacks where malicious content on searched pages can override agent instructions.
- Ingestion points: Data enters the agent context from web search and extraction tools like tavily/search-assistant and exa/search via the infsh command.
- Boundary markers: The provided prompt templates use simple text labels (e.g., 'Source 1', 'Search Results:') but do not employ robust delimiters or explicit instructions to ignore embedded commands in the retrieved data.
- Capability inventory: The skill utilizes the Bash tool to execute various subcommands of the infsh CLI to manage data flow and model interaction.
- Sanitization: There is no evidence of sanitization, filtering, or escaping of the retrieved web content before it is interpolated into LLM prompts.
Audit Metadata