courier

Pass

Audited by Gen Agent Trust Hub on Sep 4, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill describes platform features that ingest untrusted data from external sources, which creates a potential surface for indirect injection if data is not properly sanitized.
  • Ingestion points: Inbound webhooks (references/guides/webhooks.md), fetch nodes in Journeys (references/guides/journeys.md), and Segment event streams.
  • Boundary markers: Instructions advise using delimiters and templating syntax (Handlebars/Elemental) for data interpolation.
  • Capability inventory: The platform supports outbound network requests via fetch nodes and message delivery to multiple external channels.
  • Sanitization: Best practices for webhook signature verification and data masking are explicitly documented to mitigate injection risks.
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to download official binaries and libraries from vendor-controlled or well-known locations.
  • Downloads: Fetches the Courier CLI binary from the official GitHub repository and loads Web Components from the UNPKG CDN.
  • Packages: Recommends installing verified SDKs from the NPM and PyPI registries under the vendor's namespace.
  • [DYNAMIC_EXECUTION]: The Courier platform features documented in the skill include runtime logic execution.
  • Jsonnet: Support for using Jsonnet for programmatic content generation within notification templates.
  • AI Node: Ability to execute LLM prompts at runtime as part of an automation sequence to process data.
  • [COMMAND_EXECUTION]: Provides instructions for using the Courier CLI to perform ad-hoc operations, debug delivery issues, and manage workspace resources.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 4, 2026, 06:18 PM
Security Audit — agent-trust-hub — courier