What problem does it solve? When an Infrahub deployment breaks, support engineers need a consistent set of artifacts — logs, config, version info, branch state — to triage the issue. This Skill automates producing that artifact set as a single redacted diagnostic bundle, so users don't have to manually gather and sanitize dozens of files. ## Core Features & Use Cases - Guided nine-step collection workflow: Captures the problem description, connection info (URL and API token with a privacy guarantee), deployment topology (Docker Compose, Kubernetes, or local dev), and classifies the issue into one of ten categories such as git-sync, upgrade, or performance. - Read-only, infrahubctl-only probing: All instance state queries go through infrahubctl; no mutating commands, no direct database or API probes, and multi-replica task-worker logs are always collected in full. - Two-tier redaction with user review: Tier 1 automatically masks known secrets (tokens, passwords, JWTs), then Tier 2 presents IPs, hostnames, and customer strings for the user to keep or redact before the bundle is finalized. - Deterministic flag checks: Twelve pattern-based checks (e.g., CommitNotFoundError, OOM kills, default security keys) scan collected files and emit hints in flags.yml for the receiving expert. - Use Case: After a Helm upgrade leaves a branch stuck in NEED_UPGRADE_REBASE, run this Skill to gather pod logs, Helm history, and branch state into a tarball ready to hand to OpsMill support. ## Quick Start Ask the assistant to collect an Infrahub diagnostic bundle for the problem you are experiencing, then provide your Infrahub URL and optionally an API token when prompted.
