lovable-cloud-migration

Warn

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: MEDIUMCREDENTIALS_UNSAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The skill provides instructions and SQL queries to access sensitive authentication data, specifically the encrypted_password column in the auth.users table (referenced in references/migration-steps.md and references/fresh-project-path.md). This allows the agent to extract bcrypt password hashes for account migration, which exposes user credentials. The skill includes appropriate warnings for the user to handle the resulting backup files securely.- [COMMAND_EXECUTION]: The migration playbook involves the execution of powerful shell commands such as pg_restore, git clone, and rsync. It also relies on the agent's capability to perform extensive database management operations using execute_sql and apply_migration tools to rebuild schema and transfer data.- [EXTERNAL_DOWNLOADS]: The instructions recommend installing and using established third-party tools from reputable sources, including the Supabase CLI, GitHub CLI (gh), and specific versions of PostgreSQL (postgresql@18) from the Homebrew package manager. These are well-known services and are essential for the migration tasks.- [INDIRECT_PROMPT_INJECTION]: The skill has a large surface for indirect prompt injection as it ingests and processes untrusted data from the source database and repository files. The agent is instructed to review connection code and perform search-and-replace operations on database content.
  • Ingestion points: Database table contents, JSONB fields, and application source code files (supabase/functions/).
  • Boundary markers: None identified for delimiting migrated data content.
  • Capability inventory: Powerful capabilities including SQL execution, edge function deployment, and shell command access.
  • Sanitization: No specific sanitization or filtering of the migrated content is described in the instructions.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 2, 2026, 03:10 AM
Security Audit — agent-trust-hub — lovable-cloud-migration