atmos-ci
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill focuses on secure CI/CD automation patterns, emphasizing containerization and native tool usage over deprecated or opaque wrappers.
- [COMMAND_EXECUTION]: The skill facilitates the execution of Atmos and Terraform/OpenTofu commands within CI environments. This behavior is the primary and intended purpose of the skill for CI/CD orchestration and follows standard automation practices.
- [EXTERNAL_DOWNLOADS]: The instructions reference official container images (ghcr.io/cloudposse/atmos) and GitHub Actions (cloudposse/atmos/actions/github-runtime) provided by the vendor. These are trusted resources belonging to the skill author's infrastructure.
- [INDIRECT_PROMPT_INJECTION]: The skill processes pull request labels and Terraform outputs to generate matrices and job summaries. It addresses the potential attack surface by providing explicit guidance on label-based filtering (--labels=ci=auto) and Open Policy Agent (OPA) validation to prevent the automated execution of privileged components.
- [PRIVILEGE_ESCALATION]: The skill documents the minimal GitHub Actions permissions required for OIDC and reporting features (contents: read, id-token: write, statuses: write, pull-requests: write). It also includes explicit warnings about maintaining trust boundaries in fork PR scenarios.
Audit Metadata