ln-731-docker-generator

Pass

Audited by Gen Agent Trust Hub on Mar 8, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits a vulnerability surface for indirect prompt injection where untrusted input could influence generated output.
  • Ingestion points: Project metadata such as 'Project Name', 'Stack Type', and 'Versions' are ingested during Phase 1 and Phase 3 of the workflow in SKILL.md.
  • Boundary markers: Templates in the references/ directory do not utilize explicit boundary markers or delimiters to isolate substituted variables from the surrounding Dockerfile or Nginx syntax.
  • Capability inventory: The skill performs file-writing operations as described in Phase 4 of SKILL.md to create the final configuration files.
  • Sanitization: While Phase 1 includes basic validation for stack types and version ranges, there is no evidence of sanitization or escaping of project-specific strings to prevent syntax injection in generated files.
  • [EXTERNAL_DOWNLOADS]: The skill references several external resources from well-known and trusted providers.
  • Fetches the Mermaid.js library from the JSDelivr CDN for workflow visualization in diagram.html.
  • References official base images from Microsoft (mcr.microsoft.com) and standard Docker Hub registries (Node, Python, Nginx, Postgres) within the Dockerfile templates.
  • [COMMAND_EXECUTION]: The skill performs automated file system operations as part of its primary function.
  • Generates and writes multiple configuration files (Dockerfile, docker-compose.yml, nginx.conf) to the project directory based on template substitution.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 8, 2026, 11:56 PM