CONTRIBUTOR TOOL - Track CC changelog, extract new versions since last check, analyze impact on plugin (breaking changes, opportunities, deprecations). Run periodically or before releases. NOT part of the distributed plugin.

Debug Ecto constraint violations - trace triggers, check migrations, find duplicate data. Use when seeing unique_constraint, foreign_key_constraint, or check_constraint errors.

Ecto patterns: schemas, changesets, queries, migrations, transactions (Multi/transact), associations/preloads, upserts, and full-text search. Load for Repo calls, Ecto.Query, schema fields, constraints, cast/put_change, preloading, N+1 fixes, SQL injection prevention, float-for-money. Skip for Ash.Resource/Ash.Domain modules.

Scan Ecto code for N+1 query anti-patterns: Repo calls inside Enum.map/each/for loops, association access without preload, missing nested preloads. Use when user reports excessive queries in logs, wants an N+1 audit, or asks to find missing preloads. NOT for general Ecto guidance — use ecto-patterns.

OTP/BEAM patterns and core Elixir idioms — GenServer, Supervisor, Task, Registry, pattern matching, with chains, pipes, error tuples, mix tasks. Load for process design, supervision trees, language idiom refactoring, or BEAM debugging. NOT for Ecto/LiveView/Oban/Phoenix — use dedicated skills.

Fetches HexDocs documentation efficiently using WebFetch tool with automatic HTML-to-markdown conversion for context efficiency. Use this INSTEAD of direct WebFetch when looking up any Elixir library documentation on hexdocs.pm. Handles module docs, function docs, guides, and changelogs for any Hex package. Dramatically reduces token usage compared to raw HTML fetching.

Internal router for ambiguous work requests. Load when user says "add/build/implement feature", "fix/debug/investigate bug", "review/audit code", "refactor/clean up", "slow/performance", "PR comments", "research/evaluate library", or "that fixed it" — WITHOUT an explicit /phx: slash command.

Investigate bugs using the Ralph Wiggum approach: check the obvious, read errors literally.

Self-improving loop for plugin skills. Reads program.md, proposes one mutation per iteration, evaluates against deterministic scorer, keeps improvements via git, reverts failures. Targets weakest skill+dimension. Use with /loop for overnight runs.

LiveView lifecycle & real-time UI. For *_live.ex, mount, handle_event, handle_params, handle_info, assign_async, streams, PubSub, phx-change/phx-submit, uploads, JS hooks, LiveComponent, push_patch, on_mount, socket assigns. NOT for controllers, REST APIs, Ecto, or plain Phoenix.Component.

Diagnostic audit of LiveView socket assigns — inventories every assign with memory footprint estimates, detects missing temporary_assigns, finds unused assigns, and flags unbounded lists needing streams. Use when investigating LiveView memory bloat, auditing assign efficiency, or explicitly requested via /lv:assigns. NOT for general LiveView patterns or building features.

Narrow bare `rescue _ ->` / `rescue e ->` so UndefinedFunctionError, KeyError, and typos propagate instead of being swallowed. Use for auditing rescues, secure-coding review, exception review, refactoring error handling in Elixir.

Oban job processing — workers, perform/1, queues, cron scheduling, retries, unique jobs, idempotency. Oban Pro workflows/batches/process/1. Oban.Worker, Oban.Testing, assert_enqueued, perform_job, queue config, backoff. Prevents atom-key args, struct serialization, non-idempotent jobs.

Phoenix context module design: creating/splitting contexts, adding context functions, cross-context boundaries, Scope struct (1.8+), Ecto.Multi in contexts, PubSub broadcasting. Also: router pipelines, scope blocks, verified routes, plugs, controller delegation to contexts, JSON API controllers.

Project health audit across 5 areas (architecture, performance, security, tests, dependencies) using parallel specialist agents. Use quarterly, before major releases, after large refactors, or when onboarding. Also use when the user mentions "project health", "code quality overview", "how healthy is this codebase", or any request for holistic assessment.

Validate Phoenix context boundaries and dependencies using mix xref. Use when reviewing PRs, before major refactors, when contexts feel tangled, or when a module has too many cross-context dependencies. Also use when planning context splits or evaluating architecture health.

Brainstorm Elixir/Phoenix features — explore ideas, compare approaches, gather requirements. Use when vague idea, not sure how to approach, or want to discuss before plan.

Challenge mode reviews - rigorous questioning before approving changes. Use when you want thorough scrutiny of Ecto changes, LiveView events, or PR readiness.

Use after fixing any non-trivial bug, resolving a tricky issue, or when the user says "that worked" or "it is fixed". Also suggest proactively after successful /phx:review or /phx:investigate. Captures solved problems as searchable Elixir/Phoenix solution documentation for faster future debugging.

Practical examples and walkthroughs for Elixir/Phoenix patterns, including plugin workflow demonstrations. Use whenever the user asks "how do I...", "show me an example of...", "what does X look like", or needs to see a working implementation pattern for Phoenix, LiveView, Ecto, or OTP features. Also triggers when the user asks about plugin commands or wants to see what a workflow produces.