Skills
skills/1nfsh/skills/ai-rag-pipeline/Gen Agent Trust Hub

ai-rag-pipeline

Fail

Audited by Gen Agent Trust Hub on Feb 18, 2026

Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • [REMOTE_CODE_EXECUTION] (HIGH): The skill utilizes the pattern curl -fsSL https://cli.inference.sh | sh. This downloads and immediately executes a shell script from a non-trusted third-party domain, providing a direct path for arbitrary code execution on the host system.
  • [COMMAND_EXECUTION] (MEDIUM): The allowed-tools configuration specifies Bash(infsh *). This is an overly permissive policy that allows the agent to execute any command starting with infsh, which could be exploited to perform unauthorized actions or credential manipulation via infsh login if the agent's logic is subverted.
  • [PROMPT_INJECTION] (LOW): The skill exhibits a significant Indirect Prompt Injection surface. It retrieves arbitrary data from the web using search tools and interpolates the results (e.g., $SEARCH_RESULT, $CONTENT) directly into LLM prompts without using delimiters or sanitization.
  • Ingestion points: infsh app run tavily/search-assistant, infsh app run exa/search, and infsh app run tavily/extract.
  • Boundary markers: Absent. The data is directly embedded into natural language strings.
  • Capability inventory: The skill has access to Bash for command execution.
  • Sanitization: None detected. The skill relies on the LLM to 'ignore' malicious instructions embedded in the search results.
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
HIGH
Analyzed
Feb 18, 2026, 02:01 AM
Security Audit — agent-trust-hub — ai-rag-pipeline