abridge-deploy-integration
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill implements healthcare-grade infrastructure patterns, including multi-stage Docker builds and non-root user execution, which align with HIPAA best practices for container security.\n- [SAFE]: Sensitive data management is handled appropriately using placeholders for secrets and directing users to utilize cloud-native Secret Manager services rather than hardcoding credentials directly in the deployment scripts.\n- [SAFE]: Remote operations and tool usage (gcloud, aws, docker) are strictly aligned with the skill's stated purpose of cloud deployment and healthcare integration management.\n- [INDIRECT_PROMPT_INJECTION]: The skill establishes an attack surface for indirect prompt injection by interpolating external data into shell scripts and code templates without built-in sanitization.\n
- Ingestion points: External configuration variables such as GCP_PROJECT_ID, ABRIDGE_BASE_URL, and EPIC_FHIR_BASE_URL referenced in the deployment scripts and health check code.\n
- Boundary markers: Absent; the instructions do not provide delimiters or instructions for the agent to treat these inputs as untrusted data.\n
- Capability inventory: The skill uses broad capabilities including Bash(gcloud:), Bash(docker:), and Bash(aws:*) to execute the generated templates.\n
- Sanitization: Absent; the provided templates perform direct interpolation of variables into shell commands and network requests.
Audit Metadata