skills/assistant-ui/skills/update/Gen Agent Trust Hub

update

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the user to execute local shell commands such as npm ls, npm view, npx tsc, and npm run build to verify the project environment and dependency versions.
  • [REMOTE_CODE_EXECUTION]: The skill utilizes npx assistant-ui@latest to perform updates and codemods. This involves downloading and executing the framework's official CLI package from the NPM registry to apply code transformations. As these resources are owned by the skill's author, this is considered standard behavior for a migration tool.
  • [INDIRECT_PROMPT_INJECTION]: The migration process involves an inherent surface for indirect prompt injection as it processes user-controlled source code.
  • Ingestion points: The skill instructs the agent to process local source code directories (e.g., ./src) when running codemods and identifying versions.
  • Boundary markers: No specific boundary markers or "ignore embedded instructions" warnings are provided for the data ingestion process.
  • Capability inventory: The skill utilizes npx for executing codemods that perform file-system modifications and npm for environment inspection.
  • Sanitization: The skill does not describe specific sanitization or validation logic for the code transformations being applied.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 03:10 PM
Security Audit — agent-trust-hub — update