acestep-docs

Pass

Audited by Gen Agent Trust Hub on Sep 18, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: Documentation in getting-started/README.md and guides/ENVIRONMENT_SETUP.md provides standard instructions for installing the uv package manager via curl or PowerShell pipes directly into a shell or interpreter.
  • Evidence: curl -LsSf https://astral.sh/uv/install.sh | sh and powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex".
  • [EXTERNAL_DOWNLOADS]: The documentation references external resources for software installation, including a binary package from the vendor's domain and the project's official GitHub repository.
  • Evidence: Downloads from https://files.acemusic.ai/acemusic/win/ACE-Step-1.5.7z and cloning from https://github.com/ACE-Step/ACE-Step-1.5.git.
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to read and process documentation that contains executable shell commands and configuration examples. This represents a potential surface for indirect prompt injection if a downstream agent with broader tool access (such as a bash tool) interprets the documentation as a set of commands to be executed.
  • Ingestion points: Documentation files such as getting-started/README.md and guides/ENVIRONMENT_SETUP.md.
  • Boundary markers: Absent; the files contain raw commands without explicit delimiters or warnings for the AI agent.
  • Capability inventory: The skill itself is restricted to Read, Glob, and Grep tools.
  • Sanitization: Absent; the agent is instructed to read the raw content of the documentation files.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 18, 2026, 05:25 PM
Security Audit — agent-trust-hub — acestep-docs