finta-deploy-integration
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill integrates with the Finta API, which acts as an external ingestion point for potentially untrusted data. 1. Ingestion points: The health check logic in SKILL.md fetches data from the /rounds endpoint of the Finta API. 2. Boundary markers: The skill lacks explicit boundary markers or instructions for the agent to ignore embedded commands within the fetched API data, though it does include instructions for redacting sensitive payloads from logs. 3. Capability inventory: The skill is granted Read, Write, and Edit tools and provides shell commands for Docker lifecycle management. 4. Sanitization: The instructions suggest using synthetic data for staging and redacting sensitive payloads from logs, providing a partial mitigation.
- [COMMAND_EXECUTION]: The skill includes several shell commands for managing Docker containers, including building images, running containers, and performing rolling updates.
- [EXTERNAL_DOWNLOADS]: The deployment process involves fetching external resources from official registries. 1. The Dockerfile pulls the node:20-slim base image from Docker Hub. 2. The build process uses npm ci to install dependencies from the official npm registry.
Audit Metadata