circleci-automation

Pass

Audited by Gen Agent Trust Hub on Apr 7, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill instructions do not contain any malicious patterns, such as prompt injection, credential exfiltration, or obfuscated content. The author's use of a 'critical' risk label in the metadata appears to be a legitimate assessment of the high-impact nature of CI/CD automation rather than an indicator of malicious intent.
  • [EXTERNAL_DOWNLOADS]: The skill references an external service endpoint (https://rube.app/mcp) required for its operation. This is a configuration requirement for the Rube MCP tool described and does not involve the automated download or execution of untrusted scripts within the skill's own logic.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest data from external CI/CD sources, such as job artifacts and test metadata. This creates an entry point for untrusted data into the agent's context. While this is a standard risk for automation tools that process external content, no unsafe interpolation or sanitization vulnerabilities were identified in the skill's design.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 7, 2026, 02:56 PM