elixir-phoenix-plan

name	elixir-phoenix-plan
description	Plan multi-step Phoenix work, new domains, or larger multi-file changes.
metadata	{"short-description":"Plan multi-step Phoenix changes"}

Plan Elixir/Phoenix Feature

Plan a feature by spawning Elixir specialist agents, then output structured plan with checkboxes.

What Makes elixir-phoenix-plan Different from the generic plan flow

1. Keeps Elixir/Phoenix-specific research and task routing in view
2. Plans with [ecto], [liveview], [oban] task routing
3. Checks for Iron Law compliance in the plan
4. Includes mix compile/format/credo/test verification
5. Understands Phoenix context boundaries

Usage

`elixir-phoenix-plan` Add user avatars with S3 upload
`elixir-phoenix-plan` .codex/plans/notifications/reviews/notifications-review.md
`elixir-phoenix-plan` Implement notifications --depth deep
`elixir-phoenix-plan` .codex/plans/auth/plan.md --existing

Arguments

• $ARGUMENTS = Feature description, review file, or existing plan
• --depth quick|standard|deep = Planning depth (auto-detected)
• --existing = Enhance an existing plan with deeper research

Workflow

1. Gather context — File path (skip to agents), brainstorm interview.md (skip clarification), clear description, or vague
2. Clarify if vague — Ask questions ONE at a time (skip if brainstorm interview.md exists with Status: COMPLETE)
3. Detect depth — Auto-detect quick/standard/deep
4. Runtime context (Tidewave) — Gather live schemas, routes, and warnings before spawning agents (see planning-orchestrator)
5. Spawn research agents — Selective, parallel, based on need. Create a Codex task per agent for progress visibility: create a task named "{Agent} research", mark in_progress on spawn, completed when done
6. Wait for ALL agents — Do NOT proceed until all return "completed". NEVER write plan while any agent is still running
7. Breadboard (LiveView) — System map for multi-page features
8. Completeness check — MANDATORY when planning from review
9. Split decision — One plan or multiple, concrete options
10. Generate plan — Checkboxes, phased tasks, code patterns. Also create plans/{slug}/scratchpad.md for decisions and dead-ends
11. Self-check (deep only) — Three questions in Risks section
12. Present and ask — STOP, show summary, let user decide

When planning from review: Every finding must appear in the plan — either as a task OR explicitly deferred by the user.

See references/planning-workflow.md for detailed step-by-step.

--existing Mode (Deepening)

Enhances an existing plan instead of creating a new one:

1. Load plan, search .codex/solutions/ for known risks
2. Spawn SPECIALIST agents (not Explore) for thin sections. Each agent writes to .codex/plans/{slug}/research/ and returns only a 500-word summary. Same agent selection rules
3. Wait for ALL agents (mark tasks completed as each finishes)
4. Add implementation detail, resolve spikes, add verification
5. Present diff summary — NEVER delete existing tasks

Iron Laws

1. NEVER auto-start elixir-phoenix-work — Always present plan and ask
2. Research before assuming — Web-search unfamiliar tech
3. Spawn agents selectively — Only relevant, not all
4. NEVER write plan while agents still running
5. NEVER skip input findings — Every finding MUST have a task
6. Do NOT spawn hex-library-researcher for existing deps
7. Skip research when planning from review/investigation — When input is a review file or elixir-phoenix-investigate output, the findings ARE the research. Do NOT spawn agents to re-discover what the review already found. Convert findings directly to plan tasks. (Confirmed: 56-session analysis showed same findings discovered 3-4x across review→investigate→plan phases, wasting ~96K tokens)

Integration with Workflow

`elixir-phoenix-plan` {feature}  <-- YOU ARE HERE
       |
   `elixir-phoenix-plan` --existing (optional enhancement)
       |
   ASK USER -> `elixir-phoenix-work` .codex/plans/{feature}/plan.md
       |
`elixir-phoenix-review` → `elixir-phoenix-compound`

Notes

• Plans saved to .codex/plans/{slug}/plan.md
• Research reports in .codex/plans/{slug}/research/ can be deleted after

CRITICAL: After Writing the Plan

STOP. Do NOT proceed to implementation.

After writing .codex/plans/{slug}/plan.md:

1. Summarize: task count, phases, key decisions
2. Use ask the user directly with options:
   • "Start in fresh session" (recommended for 5+ tasks)
   • "Get a briefing" (elixir-phoenix-brief — interactive walkthrough)
   • "Start here"
   • "Review the plan"
   • "Adjust the plan"
3. Wait for user response. Never auto-start work.

When user selects "Start in fresh session", print:

1. Start a fresh Codex session
2. Then run one of:
   `elixir-phoenix-work` .codex/plans/{slug}/plan.md
   `elixir-phoenix-full` .codex/plans/{slug}/plan.md  (includes review + compound)

This is Iron Law #1. Violating it wastes user context.

References (DO NOT read — for human reference only)

• references/planning-workflow.md — Detailed step-by-step
• references/plan-template.md
• references/complexity-detail.md
• references/example-plan.md
• references/agent-selection.md
• references/breadboarding.md