skills/xingyu4j/skills/tsdown/Gen Agent Trust Hub

tsdown

Pass

Audited by Gen Agent Trust Hub on Mar 26, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides documentation for installing and using the tsdown package and its related ecosystem. It references legitimate tools and plugins from the Rolldown and Vite ecosystems, including typescript, unplugin-vue, and @rollup/plugin-babel. These references are standard for the tool's intended purpose of library bundling.
  • [COMMAND_EXECUTION]: The reference files include code examples demonstrating how to use lifecycle hooks (e.g., build:prepare, build:done) and programmatic APIs. These examples involve routine build operations like cleaning directories (rm -rf dist), running tests, or copying assets. These are documented as user-controlled configurations and do not represent autonomous or unauthorized command execution by the skill itself.
  • [PROMPT_INJECTION]: The skill's architecture presents a surface for indirect prompt injection as it encourages the agent to read local project configuration files (package.json, tsdown.config.ts) and source code to provide assistance. This is a common and expected interaction pattern for developer-oriented tools.
  • Ingestion points: Project configuration files (package.json, tsdown.config.ts) and source files.
  • Boundary markers: None explicitly defined in the documentation templates, relying on the agent's baseline instruction following.
  • Capability inventory: File system access and shell command execution (via user-configured build hooks or programmatic scripts).
  • Sanitization: The documentation does not explicitly detail sanitization steps for external inputs, as it focuses on developer-controlled configuration.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 26, 2026, 07:37 AM