ai-rag-pipeline

Fail

Audited by Gen Agent Trust Hub on Feb 19, 2026

Risk Level: CRITICALREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • REMOTE_CODE_EXECUTION (CRITICAL): Detected piped remote code execution via curl -fsSL https://cli.inference.sh | sh. This pattern executes scripts directly from the internet without verification, which is a severe security risk.
  • EXTERNAL_DOWNLOADS (HIGH): The skill downloads executable binaries from dist.inference.sh, a domain not included in the Trusted External Sources list. This introduces a significant supply chain vulnerability.
  • COMMAND_EXECUTION (MEDIUM): The skill requests broad shell permissions via Bash(infsh *) and constructs shell commands by interpolating variables into strings, which can lead to command injection if input is malicious.
  • PROMPT_INJECTION (LOW): The skill exhibits an Indirect Prompt Injection surface.
  • Ingestion points: Untrusted data from web searches (tavily, exa) is stored in variables like SEARCH_RESULT and CONTENT (SKILL.md).
  • Boundary markers: None. Web content is injected directly into prompt strings without delimiters.
  • Capability inventory: The skill uses infsh app run to execute remote LLM tasks and tool calls.
  • Sanitization: None. There is no evidence of escaping or validating the content retrieved from external URLs before processing.
Recommendations
  • HIGH: Downloads and executes remote code from: https://cli.inference.sh - DO NOT USE without thorough review
  • AI detected serious security threats
Audit Metadata
Risk Level
CRITICAL
Analyzed
Feb 19, 2026, 07:50 PM