stack-modernization

Pass

Audited by Gen Agent Trust Hub on Oct 7, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill leverages shell commands to perform its core modernization tasks, specifically using the Bun package manager and Git.
  • Evidence includes the use of bun add, bun remove, bun outdated, and bun run test to manage dependencies and verify project integrity.
  • [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect injection through external data sources like web search results and package manager outputs.
  • Ingestion points: The skill uses WebSearch to retrieve current package versions and migration guides from external websites (npm, GitHub).
  • Boundary markers: The instructions mandate a 'confirmation gate' where a modernization plan must be approved by the user before execution.
  • Capability inventory: The skill has authorized access to execute bun and git commands, allowing it to modify files and install packages.
  • Sanitization: The skill follows a strict incremental process, running tests after every single package change to ensure that any regressions are immediately identified and can be rolled back.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 7, 2026, 04:47 PM