fathom-deploy-integration

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process meeting transcripts, summaries, and action items retrieved from the Fathom API.
  • Ingestion points: External data enters the agent context via the Fathom API (/meetings endpoint) and incoming webhooks.
  • Boundary markers: The instructions lack specific delimiters or instructions for the agent to ignore potentially malicious content embedded within meeting transcripts.
  • Capability inventory: The skill possesses significant capabilities including file Write, Edit, and Bash(gcloud:*) for cloud infrastructure management.
  • Sanitization: No explicit sanitization, filtering, or validation logic is defined for handling the processed meeting content before it is used to influence agent decisions.
  • [EXTERNAL_DOWNLOADS]: The Docker configuration references node:20-slim, which is downloaded from Docker Hub, a well-known container registry.
  • [COMMAND_EXECUTION]: The skill utilizes several command-line tools for its primary deployment purpose:
  • Docker: Commands like docker build, docker run, and docker rm are used to manage the lifecycle of the integration container.
  • Curl: Used within the Docker health check and for verifying the service status post-deployment.
  • gcloud: The skill is configured to allow gcloud commands via the Bash tool for managing cloud deployments.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 03:47 AM
Security Audit — agent-trust-hub — fathom-deploy-integration