promote-to-prod

Pass

Audited by Gen Agent Trust Hub on Jul 30, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes standard command-line tools including the GitHub CLI (gh), yq, and git to interact with repository state and trigger CI/CD workflows. These actions are aligned with the skill's stated purpose of managing deployments.
  • [EXTERNAL_DOWNLOADS]: The skill fetches configuration data (YAML) from the official PostHog/charts repository on GitHub. This is a legitimate operation for resolving image references within the vendor's infrastructure.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by parsing external YAML data and interpolating values into shell commands. However, the risk is mitigated by the source being a controlled vendor repository and the requirement for a manual approval gate in the GitHub Actions environment before any promotion is finalized.
  • Ingestion points: state/duckgres.yaml via gh api.
  • Boundary markers: Absent.
  • Capability inventory: Shell execution of gh workflow run, gh api, and git log.
  • Sanitization: None; values are interpolated directly into shell strings.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 30, 2026, 09:16 PM
Security Audit — agent-trust-hub — promote-to-prod