skills/inf-sh/skills/infsh-cli/Gen Agent Trust Hub

infsh-cli

Fail

Audited by Gen Agent Trust Hub on Apr 4, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONDATA_EXFILTRATIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The installation process for the CLI tool involves downloading and executing a script directly from a remote URL.
  • Evidence: curl -fsSL https://cli.inference.sh | sh is listed as the primary installation method in SKILL.md and references/authentication.md.
  • [DATA_EXFILTRATION]: The CLI is designed to automatically upload local files to the inference.sh cloud when a file path is provided as an input. This could be used to send local system data to external infrastructure.
  • Evidence: SKILL.md and references/running-apps.md describe how local paths for media and data inputs are automatically uploaded during execution.
  • [COMMAND_EXECUTION]: The skill configuration allows for broad execution of the infsh utility, which includes functions for authentication, task management, and application deployment.
  • Evidence: allowed-tools: Bash(infsh *) in the YAML frontmatter of SKILL.md.
  • [EXTERNAL_DOWNLOADS]: The installer retrieves platform-specific binaries and verification checksums from an external distribution server.
  • Evidence: Documentation in SKILL.md references downloading files from dist.inference.sh.
  • [PROMPT_INJECTION]: The skill provides an attack surface for indirect prompt injection as it processes untrusted JSON inputs and local file contents through AI models without defined boundaries.
  • Ingestion points: Input JSON strings and local file paths in SKILL.md and references/running-apps.md.
  • Boundary markers: Absent.
  • Capability inventory: Bash(infsh *) which can read local files and perform network requests to a cloud provider.
  • Sanitization: Absent.
Recommendations
  • HIGH: Downloads and executes remote code from: https://cli.inference.sh - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Apr 4, 2026, 09:10 PM
Security Audit — agent-trust-hub — infsh-cli