docker-security

Pass

Audited by Gen Agent Trust Hub on Oct 8, 2026

Risk Level: SAFEPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PRIVILEGE_ESCALATION]: The skill includes a command for the CIS Docker Benchmark that requires mounting the host Docker socket (/var/run/docker.sock) and using host networking and PID namespaces. While this is a highly privileged configuration allowing container-to-host control, it is the standard and required method for the docker-bench-security tool to perform its auditing tasks.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes an image parameter that is intended to be used within shell commands for vulnerability scanning.
  • Ingestion points: The image parameter defined in the parameters section of SKILL.md.
  • Boundary markers: No specific delimiters or warnings for the agent to ignore instructions embedded in the image name are provided.
  • Capability inventory: The skill facilitates the execution of shell commands through docker, trivy, and docker scout (found in SKILL.md).
  • Sanitization: There is no explicit logic or instruction provided to sanitize the image parameter before it is interpolated into shell commands.
  • [COMMAND_EXECUTION]: The skill contains numerous shell command snippets for hardening containers, managing capabilities, and running security scanners.
  • [EXTERNAL_DOWNLOADS]: The skill references the docker/docker-bench-security image and the gcr.io/distroless/nodejs20-debian12 base image. These are downloaded from Docker's official repository and Google's container registry respectively.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 8, 2026, 09:48 PM