skills/jeremylongshore/tons-of-skills-marketplace/orchestrating-deployment-pipelines/Gen Agent Trust Hub
orchestrating-deployment-pipelines
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes deployment configurations and interacts with external build/deployment tools, creating a surface for potential instruction injection via untrusted configuration files or tool outputs.
- Ingestion points: Reads deployment configurations (example_config.yaml, pipeline_template.yaml) and processes outputs from kubectl, docker, and CI/CD platforms.
- Boundary markers: The instructions do not specify explicit delimiters or 'ignore' instructions for processing external content.
- Capability inventory: The skill uses restricted Bash tools for git, docker, and kubectl, along with filesystem Write and Edit permissions.
- Sanitization: There are no explicit sanitization mechanisms for external data before it is processed or used in shell commands.
Audit Metadata