MikaAK

Use when working with rpc_load_balancer for TODO

API wrapper patterns for the CheddarFlow project. TRIGGER when: writing or modifying API wrapper apps (apps with _api suffix), making HTTP requests to external services, working with SharedUtils.HTTP, Tesla middleware, or Finch pools. Also trigger when creating response structs with new!/1 constructors. DO NOT TRIGGER when: working with internal Elixir code that doesn't make external HTTP calls.

Authentication and authorization patterns for the CheddarFlow project. TRIGGER when: writing or modifying authentication, authorization, sessions, Auth0 integration, user access control, entitlements, or any plug/middleware that checks user identity or permissions. Also trigger when working with SessionCache, JWKS, or Ueberauth. DO NOT TRIGGER when: working with code that doesn't involve auth or user identity.

Oban background job patterns for the CheddarFlow project. TRIGGER when: writing or modifying Oban workers, background jobs, cron schedules, job queue configuration, or any code that enqueues or processes Oban jobs. Also trigger when working with Schemas.JobsRepo for job-related queries. DO NOT TRIGGER when: working with code that doesn't involve background job processing.

Caching patterns for the CheddarFlow project. TRIGGER when: writing or modifying caching code involving Redis, ETS, elixir_cache, Cachex, or any of the _cache apps. Also trigger when working with RedisLock, Cache.Redis, Cache.ETS, or cache sandbox testing. DO NOT TRIGGER when: working with code that doesn't involve caching layers.

Umbrella dependency management patterns for the CheddarFlow project. TRIGGER when: adding dependencies, creating new apps, modifying mix.exs files, working with CFX.MixHelpers, or changing the mix_helpers.exs file. Also trigger when adding tags like :root, :web, :http to app configurations. DO NOT TRIGGER when: working with code that doesn't involve dependency or project configuration.

Distributed cluster patterns for the CheddarFlow project. TRIGGER when: writing or modifying code involving cross-node communication, CFXRpc, libcluster, Phoenix.PubSub broadcasting, distributed Erlang, or node routing. Also trigger when working with SharedUtils.Cluster, CFXPubSub, or any code that needs to handle dev/test vs prod cluster differences. DO NOT TRIGGER when: working with single-node code that doesn't involve distribution.

Ecto schema, migration, and database patterns for the CheddarFlow project. TRIGGER when: writing or modifying Ecto schemas, changesets, migrations, database queries, context modules, or any code in the schemas app. Also trigger when creating new database tables, adding indexes, working with EctoShorts, QueryBuilder, TimescaleDB, or FactoryEx factories. DO NOT TRIGGER when: working with code that doesn't touch the database layer.

Use when working with caching in any Elixir project that uses the elixir_cache library. TRIGGER on any of these scenarios: (1) Adding, modifying, or debugging any cache module (`use Cache`, `_cache` suffix modules, adapter selection); (2) Choosing between adapters (ETS vs Redis vs ConCache vs PersistentTerm vs Counter vs DETS vs Agent); (3) Redis operations — hash_get/hash_set/hash_scan, json_get/json_set, sadd/smembers, scan, command, pipeline, connection pools; (4) ETS operations — update_counter, match_pattern, tab2file, rehydration, read_concurrency; (5) Strategy adapters — MultiLayer (L1/L2 caching), HashRing (distributed/consistent hashing), RefreshAhead (proactive refresh); (6) Testing caches — Cache.CaseTemplate, SandboxRegistry, sandbox?, async test isolation, cache mocking; (7) Cache patterns — get_or_create, cache invalidation, cache warming, TTL management, compression_level; (8) Infrastructure — supervision tree setup with {Cache, [modules]}, runtime config, Poolboy pools; (9) Performance — thund

Caching patterns for the elixir_cache project. TRIGGER when: writing or modifying caching code involving Redis, ETS, elixir_cache, Cachex, or any of the _cache apps. Also trigger when working with RedisLock, Cache.Redis, Cache.ETS, or cache sandbox testing. DO NOT TRIGGER when: working with code that doesn't involve caching layers.

Dependency management patterns for the elixir_cache project. TRIGGER when: adding dependencies, modifying mix.exs, or changing project configuration. DO NOT TRIGGER when: working with code that doesn't involve dependency or project configuration.

Elixir code patterns and conventions for the elixir_cache project. TRIGGER when: writing or modifying ANY Elixir code in this project, including modules, functions, GenServers, supervisors, or any .ex/.exs file. This skill should ALWAYS be read first before any other skill when working with Elixir code. DO NOT TRIGGER when: only modifying non-Elixir files like CI config or documentation.

Testing patterns for the elixir_cache project. TRIGGER when: writing or modifying tests, test support modules, test configuration, or test helper functions. Also trigger when working with Cache.Sandbox or Cache.SandboxRegistry test patterns. DO NOT TRIGGER when: writing production code that doesn't involve test infrastructure.

Use when executing deploy_ex commands to operate on running infrastructure — deploying releases, uploading to S3, running Ansible deploys, managing QA nodes, autoscaling EC2 instances, running k6 load tests, SSH-ing into instances, checking instance health, viewing logs, connecting to remote IEx, rolling back releases, restarting apps, or checking what's currently deployed. This is for RUNNING existing deploy_ex mix tasks, not for writing new code or generating infrastructure files. Triggers on: deploy to production, upload release, rollback, QA node create/destroy, scale up/down, SSH into server, app is down, check health, restart app, view logs, iex console, load test, what's deployed, ping ansible hosts. Use this even without mentioning deploy_ex — if someone wants to deploy an Elixir app, check instance status, or restart a service in this project, this skill applies.

Use when writing or modifying code inside the deploy_ex library — adding new Mix tasks, modifying existing tasks, extending AWS wrapper modules, adding EEx templates to the priv/ pipeline, writing tests for deploy_ex modules, understanding deploy_ex internals and architecture, adding config keys to DeployEx.Config, registering tasks in the TUI wizard, or adding Ansible roles. This is for CONTRIBUTING to deploy_ex itself, not for running deploy_ex commands operationally. Triggers on: add a mix task, new deploy_ex task, extend AwsAutoscaling, modify UpdateValidator, add terraform template, write test for QaNode, how does change detection work, add --dry-run flag to existing task, add config key, support Mix workspaces, priv template pipeline. Always use when modifying files in lib/deploy_ex/, lib/mix/tasks/, priv/, or test/ within deploy_ex.

Use this skill whenever the user wants to run terraform or ansible commands for AWS infrastructure. Covers: terraform plan, apply, build, init, drop, replace, refresh, state fixes; ansible setup and ansible build; database dump and restore; EBS snapshot create or delete; state bucket and lock table management; deploy_ex configuration changes; GitHub Actions CI/CD workflow installation; tearing down or destroying environments; generating PEM keys; setting up new project infrastructure from scratch; regenerating .tf files after adding releases. Do NOT use for: deploying releases with ansible.deploy, SSH access to servers, building release artifacts, or editing deploy_ex library source code.