Phase Research

---
description: Research how to implement a phase (standalone - spawns orc-phase-researcher agent)
context: inherit
user-invocable: true
allowed-tools: [Task, Read, Bash, Glob, AskUserQuestion]
---

<objective>
Research how to implement a phase. Spawns orc-phase-researcher agent with phase context.

**Use this command when:**
- You want to research without planning yet
- You want to re-research after planning is complete
- You need to investigate before deciding if a phase is feasible

**Orchestrator role:** Parse slug + phase, validate against roadmap, check existing research, gather context, spawn researcher agent, present results.

**Why subagent:** Research burns context fast (WebSearch queries, source verification). Fresh 200k context for investigation. Main context stays lean for user interaction.
</objective>

<context>
Arguments: $ARGUMENTS = `slug phase-num` (e.g., `my-app 2`)

Parse slug and phase number in step 0 before any directory lookups.
</context>

<process>

## 0. Initialize Context

Parse `$ARGUMENTS` into slug and phase number:
- `slug` = first word
- `PHASE` = second word

Load project state:
```bash
cat .claude/plans/${slug}/STATE.md
```

Find the phase directory:
```bash
ls -d .claude/plans/${slug}/phases/0${PHASE}-*/ 2>/dev/null | head -1
```

Store as `phase_dir`. Extract `phase_name` from the directory name (the part after the number and hyphen).

Check for existing research:
```bash
ls ${phase_dir}/RESEARCH.md 2>/dev/null
```

Store paths for later use:
- `state_path` = `.claude/plans/${slug}/STATE.md`
- `project_path` = `.claude/plans/${slug}/PROJECT.md`
- `roadmap_path` = `.claude/plans/${slug}/ROADMAP.md`
- `context_path` = `${phase_dir}/CONTEXT.md` (if exists)
- `research_path` = `${phase_dir}/RESEARCH.md`

## 1. Validate Phase

Read ROADMAP.md and find the phase section:
```bash
cat .claude/plans/${slug}/ROADMAP.md
```

Search for `## Phase ${PHASE}:` in the ROADMAP.md content. Extract phase number, phase name, and goal from the section.

**If phase not found:** Error and exit.
**If phase found:** Continue with extracted details.

## 2. Check Existing Research

**If RESEARCH.md exists:** Offer: 1) Update research, 2) View existing, 3) Skip. Wait for response.

**If doesn't exist:** Continue.

## 3. Gather Phase Context

Use paths from step 0 (do not inline file contents in orchestrator context):
- `project_path`
- `state_path`
- `roadmap_path`
- `context_path` (if exists)

Present summary with phase description and what files the researcher will load.

## 4. Spawn orc-phase-researcher Agent

```markdown
<research_type>
Phase Research -- investigating HOW to implement a specific phase well.
</research_type>

<key_insight>
The question is NOT "which library should I use?"

The question is: "What do I not know that I don't know?"

For this phase, discover:
- What's the established architecture pattern?
- What libraries form the standard stack?
- What problems do people commonly hit?
- What's SOTA vs what Claude's training thinks is SOTA?
- What should NOT be hand-rolled?
</key_insight>

<objective>
Research implementation approach for Phase {phase_number}: {phase_name}
Mode: ecosystem
</objective>

<files_to_read>
- {project_path} (Project vision and context)
- {context_path} (Phase context from /orc-discuss, if exists)
- {state_path} (Prior project decisions and codebase analysis)
- {roadmap_path} (Full roadmap for cross-phase awareness)
</files_to_read>

<additional_context>
**Phase description:** {phase_description}
</additional_context>

<downstream_consumer>
Your RESEARCH.md will be loaded by `orc:orc-planner` (phase planner) which uses specific sections:
- `## Standard Stack` -> Plans use these libraries
- `## Architecture Patterns` -> Task structure follows these
- `## Don't Hand-Roll` -> Tasks NEVER build custom solutions for listed problems
- `## Common Pitfalls` -> Verification steps check for these
- `## Code Examples` -> Task actions reference these patterns

Be prescriptive, not exploratory. "Use X" not "Consider X or Y."
</downstream_consumer>

<quality_gate>
Before declaring complete, verify:
- [ ] All domains investigated (not just some)
- [ ] Negative claims verified with official docs
- [ ] Multiple sources for critical claims
- [ ] Confidence levels assigned honestly
- [ ] Section names match what orc-planner expects
</quality_gate>

<output>
Write to: {phase_dir}/RESEARCH.md
</output>
```

```
Task(
  prompt=filled_prompt,
  subagent_type="orc:orc-phase-researcher",

  description="Research Phase {phase}"
)
```

## 5. Handle Agent Return

**`## RESEARCH COMPLETE`:** Display summary, offer: Plan phase, Dig deeper, Review full, Done.

**`## CHECKPOINT REACHED`:** Present to user, get response, spawn continuation.

**`## RESEARCH BLOCKED`:** Show what was attempted, offer: Add context, Try different mode, Manual.

## 6. Spawn Continuation Agent

```markdown
<objective>
Continue research for Phase {phase_number}: {phase_name}
</objective>

<prior_state>
<files_to_read>
- {phase_dir}/RESEARCH.md (Existing research)
</files_to_read>
</prior_state>

<checkpoint_response>
**Type:** {checkpoint_type}
**Response:** {user_response}
</checkpoint_response>
```

```
Task(
  prompt=continuation_prompt,
  subagent_type="orc:orc-phase-researcher",

  description="Continue research Phase {phase}"
)
```

</process>

<success_criteria>
- [ ] Phase validated against roadmap
- [ ] Existing research checked
- [ ] orc-phase-researcher spawned with context
- [ ] Checkpoints handled correctly
- [ ] User knows next steps
</success_criteria>