sts2mcp-slay-the-spire-2-ai-agent

Fail

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill requires downloading a binary mod file (STS2_MCP.dll) and a configuration file from a third-party GitHub repository (github.com/Gennadiyev/STS2MCP).
  • [REMOTE_CODE_EXECUTION]: Instructions provide a workflow to clone an external Git repository and run a Python server script (server.py) using the uv tool, which executes code from a non-whitelisted source.
  • [COMMAND_EXECUTION]: The documentation specifies shell commands for copying binaries into application directories (cp), making directories (mkdir -p), and compiling source code (dotnet build).
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted game state data (such as card names, enemy descriptions, and room types) via a local API.
  • Ingestion points: http://localhost:15526/api/v1/state via get_game_state().
  • Boundary markers: Absent in the prompt templates.
  • Capability inventory: execute_action() allows the agent to control game behavior via POST requests to the local server.
  • Sanitization: No evidence of sanitization or validation of the game state strings before they are processed by the LLM.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Sep 16, 2026, 10:43 PM
Security Audit — agent-trust-hub — sts2mcp-slay-the-spire-2-ai-agent