dockerfile-builder
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided context and requests to generate Dockerfiles, which creates a surface for potential prompt injection through untrusted data.\n
- Ingestion points: User requests for Dockerfile generation as described in the Instructions section of SKILL.md.\n
- Boundary markers: The skill instructs the agent to "understand the full context" and "always validate inputs before processing".\n
- Capability inventory: Text and code generation for DevOps and Infrastructure tasks.\n
- Sanitization: Instructions emphasize following conventions and validating inputs.\n- [EXTERNAL_DOWNLOADS]: The skill lists
dockeras a required dependency. This is a standard system requirement for building and managing container images.\n- [SAFE]: The skill contains a link to the author's repository atgithub.com/inbharatai/claude-skills, which is a legitimate vendor resource.
Audit Metadata