constructive-platform

Pass

Audited by Gen Agent Trust Hub on Aug 27, 2026

Risk Level: SAFECREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The skill documents the management and storage of sensitive credentials for database and API access.\n
  • API tokens are stored locally in ~/.cnc/config/credentials.json with restricted 0o600 permissions.\n
  • Database credentials (PGPASSWORD) and authentication tokens (CNC_TOKEN) are handled via environment variables.\n
  • The documentation notes that using administrator as a default role for anonymous access is a high-risk configuration intended only for development environments.\n- [EXTERNAL_DOWNLOADS]: The skill references external resources for platform deployment and development.\n
  • Downloads Docker images from GitHub Container Registry (ghcr.io/constructive-io/).\n
  • Installs dependencies from NPM, including makage, pgsql-test, and various @constructive-io and @pgpmjs packages.\n- [INDIRECT_PROMPT_INJECTION]: Cloud function handlers create an attack surface by processing untrusted input from HTTP requests.\n
  • Ingestion points: The handler pattern in references/cloud-functions.md accepts params and context.headers directly from incoming requests.\n
  • Boundary markers: Not explicitly defined in the provided code examples.\n
  • Capability inventory: The platform supports database queries via a GraphQL client, raw SQL execution, and CLI command execution.\n
  • Sanitization: The skill recommends using a generated typed SDK/ORM instead of raw strings to mitigate injection risks when interacting with the database.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 27, 2026, 01:06 PM
Security Audit — agent-trust-hub — constructive-platform