se-dev-plugin

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill performs several external downloads during preparation and operation:
  • Downloads the uv tool from astral.sh for Python environment management.
  • Downloads busybox64u.exe from frippery.org to provide UNIX-like utilities on Windows systems.
  • Clones plugin registries and third-party plugin source code from various GitHub repositories.
  • [REMOTE_CODE_EXECUTION]: The preparation scripts (Prepare.bat and prepare.sh) execute remote installation scripts from a remote server.
  • Evidence: powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex" in Prepare.bat.
  • Evidence: curl -LsSf https://astral.sh/uv/install.sh | env UV_NO_MODIFY_PATH=1 sh in prepare.sh.
  • [INDIRECT_PROMPT_INJECTION]: The skill clones and analyzes third-party C# source code, creating a surface for indirect prompt injection if that code contains malicious instructions.
  • Ingestion points: Third-party plugin code is cloned into the Data/Sources directory.
  • Boundary markers: None. The skill instructs the agent to read and search for patterns within the raw source files.
  • Capability inventory: The skill has capabilities to execute shell commands (via BusyBox), build code with dotnet, and perform network operations via git and uv.
  • Sanitization: No specific sanitization or filtering is performed on the downloaded code before it is processed by the agent.
  • [COMMAND_EXECUTION]: Python scripts use the subprocess module to execute git and uv commands for repository management and indexing. This behavior is consistent with the skill's primary purpose.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 09:18 AM
Security Audit — agent-trust-hub — se-dev-plugin