skills/involvex/skills/bun-win32/Gen Agent Trust Hub

bun-win32

Pass

Audited by Gen Agent Trust Hub on Sep 20, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local developer tools such as bun, dumpbin.exe, and ripgrep (rg.exe) to perform its primary function of auditing and generating code bindings. These are standard operations for a development-oriented skill.
  • [EXTERNAL_DOWNLOADS]: No external network downloads or remote script executions are performed. All dependencies are managed locally via bun install at the repository root.
  • [DATA_EXFILTRATION]: No data exfiltration patterns detected. Network access is not requested in the YAML frontmatter, and scripts focus on local file system operations within D:\repos\bun-win32 and official Windows SDK directories.
  • [PROMPT_INJECTION]: The skill contains strict 'Prohibited Actions' and 'Core Principles' in its documentation and reference files (e.g., references/agents.md) that specifically instruct the agent against fabrication, guessing, or bypassing safety/validation gates.
  • [DYNAMIC_EXECUTION]: While the skill interacts with FFI (Foreign Function Interface), it does so by generating static TypeScript bindings. It uses bun:ffi for native calls, which is the intended purpose of the repository. No unsafe eval or dynamic code loading from untrusted sources was found.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 20, 2026, 05:21 PM