gameflow
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The documentation in SKILL.md and references/cli.md provides installation commands (curl -fsSL https://install.gameflow.gg | sh and irm https://install.gameflow.gg | iex) that download and execute scripts directly in the shell. These scripts are hosted on the vendor's official domain (gameflow.gg) and are intended to install the service's command-line interface.
- [COMMAND_EXECUTION]: The skill includes shell commands and a script (scripts/check-auth.sh) that use curl to interact with the GameFlow API at https://api.gameflow.gg/v1. These commands are used for legitimate purposes such as identity verification and resource listing.
- [EXTERNAL_DOWNLOADS]: The skill references several external resources, including the vendor's documentation MCP server (https://api.gameflow.gg/mcp), API endpoints, and other related skills hosted on GitHub (GameFlowGG/gameflow-skills). These resources are all within the vendor's established infrastructure.
- [INDIRECT_PROMPT_INJECTION]: The skill connects to an external Model Context Protocol (MCP) server at https://api.gameflow.gg/mcp to fetch documentation and API schemas. This represents a potential attack surface where instructions embedded in external documentation could influence agent behavior. Ingestion points: https://api.gameflow.gg/mcp. Boundary markers: present in SKILL.md (instructions to show both names to user and ask for confirmation). Capability inventory: curl commands in scripts/check-auth.sh and CLI execution. Sanitization: not explicitly detailed for remote docs content.
- [SAFE]: The skill explicitly instructs users on secure credential management, warning against hardcoding API keys in game clients or public repositories, and recommending the use of environment variables or git-ignored configuration files.
Audit Metadata