glean-deploy-integration

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill defines an integration for Glean enterprise search, which creates a surface for indirect prompt injection by ingesting external document data into the agent context.
  • Ingestion points: External data enters the agent context via the Glean Indexing and Search APIs as implemented in the service code and health check logic.
  • Boundary markers: The instructions explicitly recommend using "mocked or sandbox fixtures" and "canary datasources" to validate behavior before production deployment.
  • Capability inventory: The skill is granted Bash, Write, and Edit tools to facilitate Docker container management and cloud deployment tasks.
  • Sanitization: Documentation provides guidance on validating document schemas and monitoring for malformed payloads using the Glean Indexing API specification.
  • [SAFE]: The Dockerfile configuration adheres to the principle of least privilege by defining a non-root system user (app) to execute the integration service.
  • [SAFE]: The skill properly handles sensitive configuration by using environment variables (e.g., GLEAN_API_KEY) and advises against using literal credentials during the build and deployment process.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 03:46 AM
Security Audit — agent-trust-hub — glean-deploy-integration