cloud-docker-registry

Pass

Audited by Gen Agent Trust Hub on Sep 20, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for fetching and analyzing content from untrusted remote container registries, which can serve as an ingestion point for malicious instructions.
  • Ingestion points: Data is pulled from external registry APIs and Docker image layers (SKILL.md).
  • Boundary markers: Instructions lack explicit markers or delimiters to differentiate registry data from the agent's core instructions.
  • Capability inventory: The agent utilizes curl, docker, crane, and trufflehog for network access and filesystem analysis (SKILL.md).
  • Sanitization: No sanitization or validation of the remote data is performed before it is processed by scanning tools or the agent's reasoning engine.
  • [COMMAND_EXECUTION]: The skill involves executing CLI tools (docker, crane, trufflehog) with arguments such as repository names and tags derived from untrusted external registry responses.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 20, 2026, 11:01 PM
Security Audit — agent-trust-hub — cloud-docker-registry