Skip to content

Profiles

github-actions[bot] edited this page Jun 11, 2026 · 5 revisions

Profiles

The first question the init wizard asks is Team profile. Your answer sets smart defaults for which rules, agents, and skills are pre-checked — you can still add or remove anything before confirming.

How profiles work

Each preset (rule, agent, skill) has a tags array listing which profiles consider it a default. When you pick a profile, every preset tagged with that profile is pre-checked in the wizard. Presets not tagged for your profile are still available — just unchecked.

When you pick a profile, every collection tagged for that profile is pre-checked in the wizard. You can add or remove any collection before confirming.

Default collections by profile

Collection Frontend Backend Full-stack DevOps Pure Infra Custom
TypeScript
Git
Security
Docs
DevOps
Next.js

Frontend

UI, design systems, accessibility

Rules

Rule Default Description
type-safety yes No any, no unguarded assertions, prefer unknown
design-system-reuse Reuse shared UI components and tokens before creating new ones

Agents

Agent Default Description
frontend-specialist yes UI implementation, design-system, i18n, a11y
test-specialist yes Test creation, refactoring, stabilization
docs-maintainer yes Documentation synchronization with code changes
code-reviewer PR review — patterns, naming, complexity, tests
a11y-specialist WCAG 2.2 A/AA audit and remediation

Skills

Skill Default Description
patterns yes Apply reusable architecture and coding patterns
docs-upkeep yes Keep docs aligned with implementation changes
workspace-hygiene yes On-demand workspace audit before commits and PRs
code-review Structured review checklist for PRs

Backend

APIs, databases, auth, services

Rules

Rule Default Description
type-safety yes No any, no unguarded assertions, prefer unknown
no-console-log Forbid console.* in production code, use logger
api-error-handling Structured error responses, never leak stack traces
security-secrets Never hardcode secrets, tokens, or credentials

Agents

Agent Default Description
backend-specialist yes API design, database patterns, error handling, auth
test-specialist yes Test creation, refactoring, stabilization
docs-maintainer yes Documentation synchronization with code changes
code-reviewer PR review — patterns, naming, complexity, tests
security-specialist Vulnerability audit, dependency scanning, secrets management

Skills

Skill Default Description
patterns yes Apply reusable architecture and coding patterns
docs-upkeep yes Keep docs aligned with implementation changes
workspace-hygiene yes On-demand workspace audit before commits and PRs
code-review Structured review checklist for PRs
api-design RESTful conventions, pagination, versioning
security-audit Dependency audit, secrets scan, OWASP patterns
migration-safety Database migration review, rollback plans

Full-stack

Frontend + backend combined

This profile is the union of Frontend and Backend defaults.

Rules

Rule Default Description
type-safety yes No any, no unguarded assertions, prefer unknown
no-console-log Forbid console.* in production code, use logger
design-system-reuse Reuse shared UI components and tokens before creating new ones
api-error-handling Structured error responses, never leak stack traces
security-secrets Never hardcode secrets, tokens, or credentials

Agents

Agent Default Description
frontend-specialist yes UI implementation, design-system, i18n, a11y
backend-specialist yes API design, database patterns, error handling, auth
test-specialist yes Test creation, refactoring, stabilization
docs-maintainer yes Documentation synchronization with code changes
code-reviewer PR review — patterns, naming, complexity, tests
a11y-specialist WCAG 2.2 A/AA audit and remediation
security-specialist Vulnerability audit, dependency scanning, secrets management

Skills

Skill Default Description
patterns yes Apply reusable architecture and coding patterns
docs-upkeep yes Keep docs aligned with implementation changes
workspace-hygiene yes On-demand workspace audit before commits and PRs
code-review Structured review checklist for PRs
api-design RESTful conventions, pagination, versioning
security-audit Dependency audit, secrets scan, OWASP patterns
migration-safety Database migration review, rollback plans

DevOps / Platform

CI/CD, containers, infrastructure-as-code

Rules

Rule Default Description
security-secrets yes Never hardcode secrets, tokens, or credentials
docker-best-practices yes Multi-stage builds, non-root users, layer caching
container-image-pinning yes Pin image versions everywhere — compose, Ansible, env vars
idempotency yes Every operation must be safe to re-run
runbook-discipline yes Keep runbooks and ADRs in sync with infra changes
ansible-conventions FQCN modules, idempotency, Jinja2 safety
jinja2-templates Safe templates with | default() and variable scope
terraform-conventions Module structure, naming, state management
kubernetes-manifests Resource limits, health checks, security context
helm-conventions Chart structure, safe defaults, secret handling
shell-script-standards set -euo pipefail, shellcheck, quoting
ci-workflow-conventions Pinned actions, OIDC, minimal permissions

Agents

Agent Default Description
docs-maintainer yes Documentation synchronization with code changes
security-specialist Vulnerability audit, dependency scanning, secrets management
infrastructure-specialist Build, CI, container, deployment config
devops-specialist CI/CD pipelines, container optimization, IaC review
ansible-specialist Ansible roles, playbooks, and Jinja2 templates
kubernetes-specialist Manifests, Helm charts, Kustomize overlays
sre-specialist SLOs, alerting, runbooks, post-mortems

Skills

Skill Default Description
docs-upkeep yes Keep docs aligned with implementation changes
workspace-hygiene yes On-demand workspace audit before commits and PRs
stack-change-review yes High-blast-radius infrastructure change review
security-audit Dependency audit, secrets scan, OWASP patterns
ci-cd-best-practices Pipeline optimization, caching strategies
infrastructure-drift-check Verify declared IaC matches deployed state before merge
rollback-plan Require tested rollback steps for every production change

Pure Infrastructure

Ansible, Kubernetes, Terraform — no application code

Use this profile for repos that contain only infrastructure: Ansible roles, Kubernetes manifests, Helm charts, Docker Compose stacks, Terraform modules. There is no application code (no JS/TS, no Python services, no CI-built binaries).

This profile is scoped to infrastructure-only repos. It defaults to git, security, docs, and devops collections — the TypeScript collection is not included since there is no application code.

Rules

Rule Default Description
security-secrets yes Never hardcode secrets, tokens, or credentials
docker-best-practices yes Multi-stage builds, non-root users, layer caching
container-image-pinning yes Pin image versions everywhere — compose, Ansible, env vars
idempotency yes Every operation must be safe to re-run
runbook-discipline yes Keep runbooks and ADRs in sync with infra changes
ansible-conventions FQCN modules, idempotency, Jinja2 safety
jinja2-templates Safe templates with | default() and variable scope
terraform-conventions Module structure, naming, state management
kubernetes-manifests Resource limits, health checks, security context
helm-conventions Chart structure, safe defaults, secret handling
shell-script-standards set -euo pipefail, shellcheck, quoting
ci-workflow-conventions Pinned actions, OIDC, minimal permissions

Agents

Agent Default Description
docs-maintainer yes Documentation synchronization with code changes
security-specialist Vulnerability audit, dependency scanning, secrets management
infrastructure-specialist Build, CI, container, deployment config
devops-specialist CI/CD pipelines, container optimization, IaC review
ansible-specialist Ansible roles, playbooks, and Jinja2 templates
kubernetes-specialist Manifests, Helm charts, Kustomize overlays
sre-specialist SLOs, alerting, runbooks, post-mortems

Skills

Skill Default Description
docs-upkeep yes Keep docs aligned with implementation changes
workspace-hygiene yes On-demand workspace audit before commits and PRs
stack-change-review yes High-blast-radius infrastructure change review
security-audit Dependency audit, secrets scan, OWASP patterns
ci-cd-best-practices Pipeline optimization, caching strategies
infrastructure-drift-check Verify declared IaC matches deployed state before merge
rollback-plan Require tested rollback steps for every production change

Custom

Pick everything individually. No collections are pre-checked. Use this when your project doesn't fit a standard profile or when you want full control.

Switching profiles

After init, switch the active profile at any time:

bluetemberg switch-profile backend

This adds any agent and skill packages belonging to the new profile's defaults that are not yet in llm/packages.json. It never removes packages automatically — instead it reports which official agent/skill packages in the manifest are not part of the new profile's defaults so you can decide whether to keep or remove them manually.

Stale detection is scoped to the official agent/skill preset catalog. Rule collections (e.g. bluetemberg-rules-typescript) and any third-party packs are never flagged — the switch cannot know which kind they are and should not touch them.

After switching, run bluetemberg sync to regenerate platform files.

Clone this wiki locally