document-findings

name	document-findings
description	Take research output and produce structured, enhanced markdown documentation with mermaid diagrams. Use when user says "write up", "document", or /document-findings. In Claude Code and Cursor, /skills lists all.

📄 Document Findings

Take research output and produce structured, enhanced markdown documentation.

📌 Rules

• Never use absolute filesystem paths in links. All links must be relative to the document they appear in
• Never invent information not present in source material. No placeholder or made-up content anywhere: no fake table rows, no TBD, no example names/dates/artifacts in Deliverables, Project tracking, Roadmap, or any other section. Only real data from source material; leave sections or tables empty when there is nothing to put.
• In Project tracking > Team: Name must be an individual person (full name). Responsibility must be exactly one of Driver, Approver, Contributor, Informed (DACI). No other roles or labels.
• When structuring from research that includes a Link tree or Sources section, preserve them in the README (under Sources and Link tree, or under Discovery > Audits for non-job doc types) so the full link tree is always present
• Always attribute content to its source
• Use mermaid diagrams for any process with 3+ steps or any hierarchy with 2+ levels
• Always include a full table of contents after the --- (or after H1 if no resume at top): link to every H2 and H3 section so readers can navigate to all parts of the document
• Use an emoji in every H2 and H3 heading in the README output

📥 Inputs

1. Project path – work/{company}/{job}/. Use work/config.md.
2. Source material – Research output from the learn skill (README.md content, including Level-0 job post and findings).
3. Document type – One of: research, analysis, solutions, project-overview

If the project path or source material is missing, ask the user before proceeding.

📤 Output

1. README – work/{company}/{job}/README.md. Single README per job per work/config.md. Include the full job post as a section in the README (see README structure). All structured output (role summary, requirements, application process, link tree, job post copy) lives in this file. When you discover or document new topics, add them to this README.

📐 Markdown Standards

📌 Headings

Use H2 (##) as top-level within documents. Reserve H1 (#) for the document title only. Use an emoji in every H2 and H3 in the README (e.g. ## 📋 Role summary, ### 📊 Project tracking).

📊 Mermaid Diagrams

Use mermaid for:

• Flowcharts - Processes, workflows, decision trees
• Sequence diagrams - Multi-step interactions
• Mind maps - Problem breakdowns, topic relationships
• Gantt charts - Timelines, phases

Example for a workflow:

```mermaid
flowchart TD
    A[Start] --> B{Decision}
    B -->|Yes| C[Action]
    B -->|No| D[Alternative]
```

Example for root cause:

```mermaid
mindmap
  root((Problem))
    Why 1
      Why 2
        Why 3
          Root Cause
```

📋 Tables

Use tables for structured comparisons, source indexes, and feature matrices. Only add rows when source material provides real data. Never add placeholder rows, TBD, example data, or made-up names/dates/artifacts.

💬 Callouts

Use blockquotes with bold labels for callouts:

> **Note:** Additional context here.

> **Warning:** Risk or caveat here.

🔗 Links and Sources

Always cite sources inline with markdown links. Group full source lists in tables.

Clickable links: Every reference to a source or URL must use markdown link syntax [title](url) so it renders as a clickable link in the README. Never write the link title and URL as separate plain text or as a bare URL.

Link everything that can be linked. Do not leave linkable items as plain text. In scope:

• Sections: Include a full table of contents after the --- divider (see README structure) linking to every H2 and H3 so readers can jump to any section.
• URLs: Use [title](url) for every source, doc, or external reference.
• People: In Project tracking > Team, give each person a document anchor (<a id="jane-doe"></a> before the name, or a heading). Anywhere that person appears (notes, Audits, Deliverables, Artifacts, etc.), link their name to that anchor ([Jane Doe](#jane-doe)).
• Tickets: When a ticket ID or URL exists, link it (e.g. [ICT-123](url) or link the ID).
• Job post: The full job post lives in the README in a 📄 Job post (full copy) section; link to it from the table of contents.
• Artifacts / docs: If an artifact or doc has a URL, link it.

When in doubt, add the link.

⚙️ Process

1. 📖 Read Source Material

Read the project README (learn skill output) from the source path. Identify Level-0 (job post) content and Findings.

1b. 📁 Merge work research (if present)

If work/{company}/{job}/work-research.md exists: read it, add an 📁 Existing work section to the README (after Application process or before Sources). Include the work-research content (existing applications, work history summary, relevant experience). Update the README table of contents to link to it. Then delete work-research.md.

1c. 📄 Place connection message, cover letter, and resume at top (if present)

If work/{company}/{job}/connection-message.md or work/{company}/{job}/recruiting-contacts.md exists: place ✉️ Connection message / Recruiter outreach at the very top of the README. Include: (1) for each contact from recruiting-contacts.md: name, LinkedIn, and tastes/interests; (2) a line "Use this when connecting on LinkedIn:" and the draft connection message from connection-message.md (or the draft from recruiting-contacts.md if connection-message.md is absent). Then add a horizontal rule ---. If work/{company}/{job}/cover-letter.md or work/{company}/{job}/resume-alignment.md exists: place cover letter next (full content of cover-letter.md if present). Then add ---. Then place the resume (full content of resume-alignment.md if present). Then add a second ---. Then ensure the document has a table of contents right after that linking to every H2 and H3 in the rest of the document (including ✉️ Connection message when present). All content below the second --- (role summary, requirements, existing work, sources, link tree, etc.) stays under that. Use emoji in every section header below the ---. Then delete resume-alignment.md, cover-letter.md, connection-message.md, and recruiting-contacts.md after merging.

2. 📑 Organize by Sections

Map content into the README sections based on document type.

3. ✍️ Write README

• If connection message and/or recruiting contacts were merged, the README starts with ✉️ Connection message / Recruiter outreach (contacts table + "Use this when connecting on LinkedIn:" + draft message), then ---. Then cover letter (if any), then ---, then resume (if any), then ---, then title and navigation.
• If only cover letter and/or resume were placed at top, the README starts with cover letter (if any), then ---, then resume (if any), then ---, then title and navigation.
• After the second --- (or at start if no cover letter/resume): add a clear title (H1) and a full table of contents linking to every H2 and H3 so readers can navigate to all sections (including ✉️ Connection message when present).
• Use an emoji in every H2 and H3 (e.g. ## 📋 Role summary, ## ✅ Requirements).
• Use mermaid diagrams where relationships or flows exist.
• Use tables where comparisons or indexes exist (e.g. requirements list).
• Keep paragraphs short (3-4 sentences max).
• Cite sources inline.

5. 📐 README structure

For job applications (when source is learn output for a job post), use this structure. Order: (1) ✉️ Connection message / Recruiter outreach at very top when connection-message.md and/or recruiting-contacts.md was merged, (2) first ---, (3) Cover letter if cover-letter.md was merged, (4) second ---, (5) Resume if resume-alignment.md was merged, (6) third ---, (7) H1 title, (8) full table of contents linking to every H2 and H3 below (including ✉️ Connection message when present), (9) sections with emoji in every header. Only include sections for which source material provides real data; no placeholders or made-up content.

[Optional: ✉️ Connection message / Recruiter outreach – contacts table, then "Use this when connecting on LinkedIn:" and draft message]

---

[Optional: full cover letter from cover-letter.md when present]

---

[Optional: full resume from resume-alignment.md when present]

---

# {Role Title} at {Company}

- [✉️ Connection message / Recruiter outreach](#-connection-message--recruiter-outreach) (when present)
- [📋 Role summary](#-role-summary)
- [✅ Requirements](#requirements)
- [⏱ Years of experience required](#-years-of-experience-required)
- [📌 Responsibilities](#responsibilities)
- [📮 Application process](#application-process)
- [📄 Job post (full copy)](#-job-post-full-copy)
- [📁 Existing work](#existing-work)
- [🔗 Sources](#sources)
- [🌐 Link tree](#link-tree)
(Include only sections that exist; link to every H2 and H3 in the doc.)

---
## 📋 Role summary
One-line role, level, location, and key focus from the job post.

## ✅ Requirements
Bullets or table of must-haves and nice-to-haves from the post.

## ⏱ Years of experience required
One clear line from the job post (e.g. "Total: 12+ years. Leadership: 5+ years." or "Not specified."). Source: learn output Findings > Years of experience required. This drives resume experience cutoff in resume-alignment.

## 📌 Responsibilities
Key responsibilities and focus areas.

## 📮 Application process
How to apply, link, deadlines, any instructions.

## 📄 Job post (full copy)
Full job listing content (Level-0 from learn output). Preserve formatting for reference when drafting applications.


Past applications (company | job | key points), work history summary, and relevant experience from work/ refs. Only when work-research was merged.


Table of crawled links (URL | Depth | Title). Every URL as []().


Traversal map from learn output. Every item as []().

For other document types (analysis, solutions, project-overview), use the structure below. Full navigation at the top: after the H1 title, add a table of contents that links to every H2 and H3. Use an emoji in every H2 and H3. Section headers below describe what goes in each place when source material provides it; do not populate with placeholder or made-up data.

# {Project Name}

- [🔍 Discovery](#discovery)
  - [📊 Project tracking](#project-tracking)
  - [📋 Audits](#audits)
  - [👥 Users + Needs](#users--needs)
- [💡 Exploration](#exploration)
  - [💡 Ideation](#ideation)
  - [📎 Artifacts](#artifacts)
  - [✅ Validation](#validation)
- [🚀 Go to market](#go-to-market)
  - [📦 Deliverables](#deliverables)
  - [📈 Performance](#performance)
  - [🔄 Next version](#next-version)
(Include only sections that exist; link every H2 and H3.)

---
## 🔍 Discovery

### 📊 Project tracking
- **Team:** One row per person. Name = individual contributor (person's full name, not team or space). Responsibility = exactly one of: **Driver**, **Approver**, **Contributor**, **Informed** (DACI only; no other values). Example: ticket assignee (Ryan Allen on the UX ticket) is **Contributor**; the person who assigned the ticket is at least **Informed** (and may also be Driver, Approver, or Contributor).
-  Project + ticket | Due date.
  Name | Current state | Desired state.


Notes, current-state review, competitor review. Track artifacts in Exploration > Artifacts. When source material includes a  or  section (from learn output), preserve it here: include a "Link tree" subsection with the full traversal map and a Sources table when useful so every crawled link remains in the README.


Users raw & encoded needs (User | Time/date | Verbatim | Encoded needs). Sorted needs. Refined problem statement (In which way might we enable ${user} to solve ${mainNeed1} & ${mainNeed2}, to ${userGoal} & ${businessGoal}?).

---



Hypotheses (If | then | due to).


Artifacts table (Date | Creator | Artifact). Drawings, surveys, flow charts, prototypes, UI, code repos.  physical design files (Figma, Sketch); link or list with Creator and Artifact.


Test plan (general, users, goals). User testing results (Version | KPI 1 | KPI 2).

---



Designs for engineers (Date | Creator | Artifact). Production roadmap (Project name | Due date | Status | Person assigned).


Quality assurance. Production testing (Version | KPI 1 | KPI 2).


Learnings. Recommendations. Links to new docs.