blumeops/docs/tutorials/ai-assistance-guide.md
Erich Blume c130f72204 Update all docs titles to human-readable and delete test cards
Titles now use proper casing (e.g. "Borgmatic" instead of "borgmatic",
"AI Assistance Guide" instead of "ai-assistance-guide") and match file
stems so wiki-links work without alternate anchor text. Also removes
the title-test-alpha/beta cards and their reference index entry.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-07 21:40:22 -08:00

4.5 KiB

title tags
AI Assistance Guide
tutorials
ai

AI Assistance Guide

Audiences: AI, Owner

This guide provides context for AI agents (like Claude Code) assisting with BlumeOps operations, and helps Erich understand how to work effectively with AI assistance.

Critical Rules

These are non-negotiable for AI agents working in this repo:

  1. Always use --context=minikube-indri with kubectl - Work contexts exist that must never be touched
  2. Run mise run zk-docs at session start - Review current infrastructure state
  3. Never commit secrets - The repo is public at github.com/eblume/blumeops
  4. Wait for user review before deploying - Create PRs, don't auto-deploy
  5. Never merge PRs without explicit request - The user merges after review

Full rules are in the repo's CLAUDE.md.

Workflow Conventions

Feature Branches

All work happens on feature branches:

git checkout main && git pull
git checkout -b feature/descriptive-name
# ... make changes ...
git commit -m "Description"

Pull Requests

Use the forge's tea CLI:

tea pr create --title "Title" --description "$(cat <<'EOF'
## Summary
- Change 1
- Change 2

## Deployment and Testing
- [ ] Test step
EOF
)"

Changelog Fragments

Add a fragment for user-visible changes:

echo "Description" > docs/changelog.d/branch-name.feature.md

Types (file suffix): .feature, .bugfix, .infra, .doc, .ai, .misc

Use simple wiki-links without alternate text or extra spaces:

  • Prefer [[borgmatic]] over [[borgmatic|Borgmatic]]
  • Only use alternate text when grammatically warranted (e.g., [[cluster|Kubernetes]] reads better than [[cluster]])
  • No spaces around the pipe: [[path|Text]] not [[ path|Text ]]

When editing documentation, rewrite links to follow this convention as you encounter them.

Service Locations

Understanding where services run helps target changes correctly:

Location Services Management
indri (native) Forgejo, Zot, Jellyfin, Caddy Ansible
[[cluster Kubernetes]] Everything else

Mise Tasks

BlumeOps operations are driven by mise tasks. Run mise tasks to list all available tasks.

Task When to Use
zk-docs At session start - review infrastructure documentation
provision-indri Deploy changes to indri-hosted services via Ansible
services-check After deployments - verify all services are healthy
pr-comments Check unresolved PR comments during review
blumeops-tasks Find pending tasks from Todoist
container-list View available container images and tags
container-tag-and-release Release a new container image version
dns-preview Preview DNS changes before applying
dns-up Apply DNS changes via Pulumi
tailnet-preview Preview Tailscale ACL changes
tailnet-up Apply Tailscale ACL changes via Pulumi
docs-check-links Validate wiki-links in documentation (includes orphan detection)
docs-check-index Check every doc is referenced in its category index
docs-check-filenames Check for duplicate doc filenames
docs-review-stale Report docs by last-modified date, highlight stale ones
docs-review-tags Print frontmatter tag inventory across all docs
docs-review-random Select a random doc card for review
indri-runner-logs View Forgejo workflow logs from local runner

For ArgoCD operations, use the argocd CLI directly:

  • argocd app diff <service> - Preview changes
  • argocd app sync <service> - Deploy changes

Reference Navigation

For AI agents building context:

  • reference - Entry point for technical details
  • hosts - What hardware exists
  • apps - What's deployed in Kubernetes
  • routing - How services are exposed

Credential Access

Credentials live in 1Password. Never retrieve them directly - use existing patterns:

  • Ansible pre_tasks gather secrets at playbook start
  • external-secrets syncs to Kubernetes
  • Scripts use op CLI with user biometric prompts

Common Pitfalls

Pitfall Correct Approach
Missing kubectl context Always add --context=minikube-indri
Deploying without review Create PR first, wait for user approval
Re-explaining reference material Link to reference cards instead
Committing to main Use feature branches
Guessing at credentials Ask user or check 1Password patterns