finish-setup

Pass

Audited by Gen Agent Trust Hub on Jun 28, 2026

Risk Level: SAFECREDENTIALS_UNSAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The skill accesses sensitive environment configuration files (.env, .env.local) to verify the presence of integration keys. It also writes new configuration values (such as Stripe price IDs) to these files during the setup process.
  • [COMMAND_EXECUTION]: The skill is instructed to run local project scripts, specifically database migration commands like db:migrate or their equivalents, based on the detected project structure.
  • [PROMPT_INJECTION]: An indirect prompt injection surface is present as the skill parses and processes untrusted project data including billing plan definitions, migration files, and environment templates to drive its actions.
  • Ingestion points: lib/billing/plans.ts, db/migrations/, and .env.example.
  • Boundary markers: None identified.
  • Capability inventory: Local shell command execution and API interactions via Stripe, GitHub, Resend, and PostHog MCP servers.
  • Sanitization: No validation or sanitization of the content from the ingested project files is specified.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 28, 2026, 12:34 AM
Security Audit — agent-trust-hub — finish-setup