# 204 Synthesize 1ba4b73e

> Synthesize discovery findings into a consolidated analysis document with proposed tickets

- Skill: `tools-only/204-synthesize-1ba4b73e` (Agent Skill, multi-file: 3 files)
- Install (CLI): `npx skillmds@latest add tools-only/204-synthesize-1ba4b73e`
- Raw SKILL.md: https://api.skillmd.com/api/skills/tools-only/204-synthesize-1ba4b73e/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Product & Planning
- Author: tools-only (https://skillmd.com/u/tools-only)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/tools-only/204-synthesize-1ba4b73e

---


# Synthesize Discovery Findings

**Arguments:** $ARGUMENTS

---

## Argument Parsing

Parse the arguments to extract:
- `{Source_URLs}` - One or more Confluence document URLs (space-separated)
- `{Target_Epic}` - Optional target implementation epic (via `--target=CC-XX`)

**Examples:**
- `/synthesize-discovery https://confluence/doc1` → Single source, auto-detect target epics
- `/synthesize-discovery https://confluence/doc1 https://confluence/doc2` → Multiple sources
- `/synthesize-discovery https://confluence/doc1 --target=CC-62` → Single source, specific target epic
- `/synthesize-discovery doc1-url doc2-url doc3-url --target=CC-62` → Multiple sources, specific target

**Supported Source Types:**
- Discovery Documents (from `/create-epic-discovery`)
- Research findings documents
- Interview summaries
- Technical spike reports
- Competitive analysis documents
- Any Confluence page with structured findings

---

## Workflow Chain

```
/create-epic-discovery <epic-key>  → Discovery Document
         ↓
[Manual research, interviews, experiments]
         ↓
/synthesize-discovery <doc-urls...>  → Synthesis Document (YOU ARE HERE)
         ↓
/approve-synthesis <synthesis-url>  → Creates Jira Tickets
         ↓
/create-implementation-plan <overview-doc>  → Implementation planning continues
```

---

## Phase 0: Source Retrieval

1. **Fetch all source documents** from the provided URLs

2. **Extract from each source:**
   - Document type (discovery doc, research findings, spike report, etc.)
   - Key findings and insights
   - Validated/invalidated hypotheses
   - Answered research questions
   - Remaining unknowns
   - Recommendations

3. **Identify target implementation epics:**
   - If `--target` specified, use that epic
   - Otherwise, extract from "Target Implementation Epics" sections
   - If none found, ask user to specify

4. **FAILURE CONDITION - Missing Information:**

   If sources cannot be fetched or no target epics identified:

   **STOP and prompt the user:**
   ```
   I was unable to retrieve required information.

   Sources Retrieved: [count]/[total]
   Failed Sources: [list URLs that failed]
   Target Epics Found: [list or "None"]

   Please provide:
   1. Confirm source URLs are accessible
   2. Target implementation epic(s): [e.g., CC-62, CC-60]
   ```

   **DO NOT PROCEED** until confirmed.

5. **MANDATORY CHECKPOINT - Sources Confirmation:**

   ```
   Please confirm before I proceed:

   Sources to Synthesize:
   1. [URL] - [Document Title] ([type])
   2. [URL] - [Document Title] ([type])
   ...

   Target Implementation Epic(s): [list]

   Is this correct? (Yes / No / Correct)
   ```

   **DO NOT PROCEED** without explicit user confirmation.

---

## Phase 1: Cross-Source Analysis

1. **Consolidate findings across all sources:**
   - Merge hypothesis validation results
   - Combine research question answers
   - Identify consistent themes
   - Note contradictions or conflicts

2. **Create findings inventory:**

   | Finding ID | Source(s) | Category | Finding | Confidence | Actionable |
   |------------|-----------|----------|---------|------------|------------|
   | F1 | Doc1, Doc2 | User Need | Users want X | High | Yes |
   | F2 | Doc1 | Technical | API supports Y | Med | Yes |
   | F3 | Doc3 | Business | ROI is Z | Low | Needs validation |

3. **Identify patterns:**
   - Recurring themes across sources
   - Convergent conclusions
   - Divergent opinions requiring resolution

4. **Catalog remaining unknowns:**
   - Questions still unanswered
   - New questions that emerged
   - Areas needing further research

---

## Phase 2: Recommendation Generation

1. **Generate feature recommendations:**
   For each actionable finding, determine:
   - Should this become a ticket?
   - What type of work is it? (Feature, Enhancement, Spike, Research)
   - Which implementation epic does it belong to?
   - What's the priority based on findings?
   - Are there dependencies on other recommendations?

2. **Create recommendation matrix:**

   | Rec ID | Based On | Title | Type | Target Epic | Priority | Dependencies |
   |--------|----------|-------|------|-------------|----------|--------------|
   | R1 | F1, F2 | Implement X feature | Story | CC-62 | High | None |
   | R2 | F3 | Validate ROI assumption | Spike | CC-60 | Med | R1 |
   | R3 | F1 | Design user flow for X | Task | CC-62 | High | None |

3. **Identify scope decisions needed:**
   - Features that could go multiple directions
   - Trade-offs requiring stakeholder input
   - MVP vs future phase decisions

---

## Phase 3: Synthesis Document Creation

Create a comprehensive synthesis document:

### 1. Synthesis Overview
- **Discovery Epic:** Link to original discovery epic
- **Sources Analyzed:** List all source documents with links
- **Synthesis Date:** When this synthesis was created
- **Target Implementation Epics:** Links to target epics

### 2. Executive Summary
- **Key Insight:** One-paragraph summary of most important discovery
- **Recommendation:** High-level direction based on findings
- **Confidence Level:** Overall confidence in recommendations
- **Major Decisions Needed:** List any blocking decisions

### 3. Consolidated Findings

#### Validated Hypotheses
| ID | Original Hypothesis | Validation Status | Evidence | Implications |
|----|---------------------|-------------------|----------|--------------|
| H1 | [statement] | Validated | [sources] | [what this means] |
| H2 | [statement] | Partially Validated | [sources] | [caveats] |
| H3 | [statement] | Invalidated | [sources] | [pivot needed] |

#### Research Questions Answered
| ID | Question | Answer | Confidence | Source(s) |
|----|----------|--------|------------|-----------|
| Q1 | [question] | [answer] | High/Med/Low | [docs] |

#### Key Insights
Narrative summary of the most important learnings, organized by theme:

**User/Customer Insights:**
- [Insight with supporting evidence]

**Technical Insights:**
- [Insight with supporting evidence]

**Business Insights:**
- [Insight with supporting evidence]

### 4. Remaining Unknowns

| ID | Unknown | Impact | Recommendation | Priority |
|----|---------|--------|----------------|----------|
| U1 | [what we don't know] | [why it matters] | [next step] | High/Med/Low |

### 5. Recommendations by Epic

For each target implementation epic:

#### {Epic_Key}: {Epic_Title}

**Recommended Tickets:**

| # | Title | Type | Priority | Story Points | Dependencies | Based On |
|---|-------|------|----------|--------------|--------------|----------|
| 1 | [title] | Story | High | 5 | None | F1, F2 |
| 2 | [title] | Task | Med | 3 | #1 | F3 |

**Scope Recommendations:**
- What should be included in MVP
- What should be deferred to future phases
- What should NOT be built based on findings

**Risk Adjustments:**
- New risks identified from discovery
- Updated risk levels based on findings
- Mitigation strategies learned

### 6. Decision Log

Decisions that need to be made before implementation:

| Decision | Options | Recommendation | Rationale | Owner | Deadline |
|----------|---------|----------------|-----------|-------|----------|
| [decision] | A, B, C | B | [why] | [who] | [when] |

### 6a. Blocking Decisions

**Status:** [Pending Resolution / All Resolved]

Decisions that MUST be resolved before tickets can be created:

| ID | Decision | Options | Status | Resolution | Resolved By | Date |
|----|----------|---------|--------|------------|-------------|------|
| D1 | [decision question] | A, B, C | Pending | - | - | - |

**Impact:** Tickets [T1, T3, T5] are blocked until all decisions are resolved.

*Note: Use `/approve-synthesis` to resolve these decisions and create tickets.*

### 7. Source Cross-Reference

Map findings back to sources:

| Source | Key Contributions | Findings Referenced |
|--------|-------------------|---------------------|
| [Doc1 URL] | [what this doc contributed] | F1, F2, F5 |
| [Doc2 URL] | [what this doc contributed] | F3, F4 |

### 8. Appendix: Full Findings Inventory

Complete list of all findings extracted from sources with full detail.

### 9. Proposed Tickets Data

**Status:** Pending Approval
**Last Updated:** [timestamp]

<!-- PROPOSED_TICKETS_START -->
```json
{
  "version": "1.0",
  "synthesis_url": "[this document URL]",
  "discovery_epic": "[Discovery Epic Key]",
  "blocking_decisions": [
    {
      "id": "D1",
      "question": "[decision question]",
      "options": ["A", "B", "C"],
      "status": "pending",
      "resolution": null,
      "blocks_tickets": ["T1", "T3"]
    }
  ],
  "proposed_tickets": [
    {
      "id": "T1",
      "title": "[Ticket title]",
      "type": "Story",
      "target_epic": "CC-62",
      "priority": "High",
      "story_points": 5,
      "dependencies": [],
      "blocked_by_decisions": ["D1"],
      "based_on_findings": ["F1", "F2"],
      "description": "[Full ticket description]",
      "acceptance_criteria": [
        "Criterion 1",
        "Criterion 2"
      ],
      "notes_from_discovery": "[Relevant insights]"
    }
  ],
  "approval_status": "pending",
  "approved_by": null,
  "approved_date": null
}
```
<!-- PROPOSED_TICKETS_END -->

*This section is machine-readable and used by `/approve-synthesis` command.*

---

## Phase 4: Review & Approval

**MANDATORY CHECKPOINT - Synthesis Review:**

```
## Synthesis Document Preview

[Full synthesis document content]

---

Summary:
- Sources Analyzed: [count]
- Findings Extracted: [count]
- Hypotheses Validated: [count] / Invalidated: [count]
- Proposed Tickets: [count]
- Blocking Decisions: [count] pending
- Target Epics: [list]

Ready to publish synthesis document? (Yes / No / Modify)
```

- **Yes** → Publish synthesis document to Confluence
- **No** → Ask what changes are needed
- **Modify** → User provides feedback, regenerate

**DO NOT PROCEED** without explicit approval.

---

## Phase 5: Publish & Output

1. **Publish synthesis document** to Confluence:
   - Location: `/epics/Discovery/{Discovery_Epic_Key}/Synthesis/`
   - Title: "{Discovery_Epic_Key} Synthesis - [Date]"

2. **Update discovery document** with link to synthesis

3. **Update target epic(s)** with:
   - Link to synthesis document
   - Summary of proposed tickets (not created yet)

---

## Output

**When complete, you MUST provide:**

```
## Discovery Synthesis Complete!

**Synthesis Document:** {Synthesis_Document_URL}

### Sources Analyzed
1. [Doc1 Title]({url})
2. [Doc2 Title]({url})
...

### Summary
- Total Findings: [count]
- Validated Hypotheses: [count]
- Invalidated Hypotheses: [count]
- Remaining Unknowns: [count]

### Proposed Tickets Summary

**Total Proposed:** [count] tickets, [sum] story points
**Blocking Decisions:** [count] pending

#### By Epic:
- {Epic_Key_1}: [count] tickets ([sum] points)
- {Epic_Key_2}: [count] tickets ([sum] points)

### Blocking Decisions
1. [D1: Decision needing resolution]
2. [D2: Decision needing resolution]

### Next Steps
1. Review the synthesis document
2. Resolve blocking decisions
3. Run ticket approval:

/approve-synthesis {Synthesis_Document_URL}
```

---

## Failure Conditions

| Condition | Action |
|-----------|--------|
| Source URL not accessible | List failed URLs, ask for alternatives |
| No target epic specified or found | Ask user to specify target epic(s) |
| Conflicting findings across sources | Document conflicts, ask user to resolve |
| Confluence publish fails | Provide document content for manual publishing |

---

## Agent Delegation

**Before synthesis:**

- **Document Analysis:** Use Explore agents to analyze each source document in parallel
- **Pattern Recognition:** Identify themes and patterns across sources
- **Cross-Reference:** Map findings to their sources

**During document creation:**

- **Validation:** Verify target epics exist and are accessible
- **Decision Identification:** Flag decisions that will block ticket creation