buildkite-pipelines
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides instructional content and code examples for Buildkite CI/CD configuration. All scripts and commands are within the scope of legitimate CI/CD operations and provided by the vendor.
- [INDIRECT_PROMPT_INJECTION]: The skill describes methods for ingesting and processing data from various sources, identifying a common CI/CD attack surface.
- Ingestion points: references/advanced-patterns.md (git diff processing), references/dynamic-pipeline-patterns.md (artifact and metadata retrieval).
- Boundary markers: Documentation explicitly details variable escaping (e.g., $$) to control interpolation during pipeline uploads.
- Capability inventory: The generated pipelines execute shell commands as part of the standard CI/CD workflow.
- Sanitization: The skill recommends best practices like disabling fork builds for sensitive pipelines and using manual approval gates for untrusted contributions.
- [SAFE]: External downloads and package references target official vendor repositories or standard registries, aligning with secure supply chain practices.
Audit Metadata