brainstorming
Structured design dialogue that validates ideas before implementation begins.
This skill should be used when enhancing FHIR meeting minutes by synthesizing transcript discussion into Confluence pages, capturing reasoning and trade-offs with XML DOM manipulation
Full skill instructions
This skill enhances FHIR Infrastructure meeting minutes by adding detailed discussion context from transcripts to Confluence pages. The process emphasizes content synthesis over transcription and XML DOM manipulation over string operations.
This skill should be used when:
This skill requires:
.env file: Contains API credentials for Confluence (and optionally JIRA)To create effective enhanced minutes:
To avoid poor quality output:
To build HTML correctly:
ONE EXCEPTION: Post-process lxml output for XHTML self-closing tags (br/hr/col). This is the only acceptable string manipulation.
To download the current Confluence page, use scripts/confluence_api.py:
# From URL (automatically extracts page ID)
python scripts/confluence_api.py "https://confluence.hl7.org/display/FHIR/Minutes"
# From page ID directly
python scripts/confluence_api.py 123456
Output: current_page.json with HTML and version number
API endpoint: GET /rest/api/content/{pageId}?expand=body.storage,version
To gather background context on JIRA issues, use scripts/jira_api.py:
# Download issue JSON
python scripts/jira_api.py FHIR-12345
# Create markdown background summary
python scripts/jira_api.py FHIR-12345 --background
Output:
FHIR-12345.json - Full issue dataFHIR-12345-background.md - Background summaryAPI endpoint: GET /rest/api/2/issue/{issueKey} (works unauthenticated for HL7)
CRITICAL: This provides background context ONLY. Never claim JIRA content was discussed today.
To understand what actually happened in the meeting:
DO NOT SKIP THIS STEP. This is where understanding of actual meeting discussion occurs.
To preserve JIRA widgets from the original page, use scripts/extract_jira_widgets.py:
python scripts/extract_jira_widgets.py current_page.json
Output: jira_widgets.json - Mapping of issue keys to widget HTML
What this does:
<ac:structured-macro ac:name="jira"> elementsIn Python:
from extract_jira_widgets import extract_jira_widgets, get_widget_element
widgets = extract_jira_widgets(html_content)
widget_elem = get_widget_element(widgets['FHIR-12345'])
To extract substance from the transcript, analyze each issue discussed by identifying:
Time investment: Allocate 15-30 minutes per issue. Going faster indicates insufficient depth.
Create working notes for each issue capturing these elements. Avoid rushing this phase.
To create final content, draft in Markdown first, then convert to HTML. For each issue:
## FHIR-12345
### Discussion
[Opening paragraph: What was discussed? What sparked it?]
[Body paragraphs: Synthesize key points, perspectives, and reasoning]
The discussion centered on [main topic]. [Participants/The team]
emphasized [key perspective], noting that [reasoning]. However,
[alternative perspective] was also raised regarding [aspect].
Several approaches were considered:
- Approach A: [benefit] but [drawback]
- Approach B: [benefit] but [drawback]
The team decided to [decision], based primarily on [reasoning].
This approach was favored because [why it addresses concerns].
[Note any remaining concerns or open issues].
Voice and Tense:
Focus on WHY:
Synthesis, Not Transcription:
Length: Aim for 2-4 substantive paragraphs per issue. One-paragraph summaries lack sufficient depth.
To convert markdown drafts to HTML and assemble the complete document, use scripts/markdown_to_html.py and scripts/assemble_minutes.py:
#!/usr/bin/env python3
import sys
sys.path.insert(0, 'scripts')
from confluence_api import download_page
from extract_jira_widgets import extract_jira_widgets, get_widget_element
from markdown_to_html import parse_simple_markdown, create_issue_section
from assemble_minutes import build_complete_minutes
import json
# Load data
with open('current_page.json') as f:
page_data = json.load(f)
original_html = page_data['body']['storage']['value']
widgets = extract_jira_widgets(original_html)
# Build enhanced sections
enhanced_sections = []
for issue_key in ['FHIR-12345', 'FHIR-12346']: # Issue keys
# Read markdown draft
with open(f'{issue_key}-draft.md') as f:
draft_md = f.read()
# Parse to lxml elements
discussion_elems = parse_simple_markdown(draft_md)
# Get JIRA widget
widget = get_widget_element(widgets[issue_key])
# Create section
section = create_issue_section(issue_key, widget, discussion_elems)
enhanced_sections.extend(section)
# Assemble complete document
final_html = build_complete_minutes('current_page.json', enhanced_sections)
# Save
with open('enhanced_minutes.html', 'w') as f:
f.write(final_html)
Key functions available in scripts:
parse_simple_markdown(text) - Convert markdown to lxml elementscreate_heading(level, text) - Create h1-h6 elementcreate_paragraph(text) - Create p elementcreate_list(items, ordered=False) - Create ul/ol elementcreate_issue_section(key, widget, discussion_elems) - Complete issue sectionassemble_enhanced_minutes(original_html, enhanced_sections) - Preserve structurepost_process_for_xhtml(html) - Fix br/hr/col tags (the ONE exception)To upload the enhanced minutes, use scripts/upload_to_confluence.py:
python scripts/upload_to_confluence.py \
<page_id> \
enhanced_minutes.html \
current_page.json
Requires environment: CONFLUENCE_URL, CONFLUENCE_USER, CONFLUENCE_TOKEN
In Python:
from upload_to_confluence import upload_from_file
response = upload_from_file(
page_id='123456',
html_file='enhanced_minutes.html',
current_version_file='current_page.json'
)
API endpoint: PUT /rest/api/content/{pageId}
To verify success, check output for HTTP 200 and new version number.
Expected time allocation per phase:
Total: 2.5-4.5 hours
Most time should be in Phases 3-4 (content work). Spending more time on technical work than content work indicates incorrect prioritization.
Shallow summaries: "We discussed FHIR-12345 and decided to implement it."
Conflating sources: Using JIRA decisions as if discussed today
Over-quoting: Verbatim transcript with timestamps
Missing WHY: Stating decisions without reasoning
String manipulation for HTML: Using regex/concat to build HTML
Forgetting deepcopy(): append() moves elements, not copies
from copy import deepcopy and use deepcopy(elem) when preservingRemoving JIRA widgets: Building content without preserving widgets
Excessive post-processing: Using regex to fix structural issues
To evaluate completion quality:
All scripts in scripts/ directory support:
.envDownload Confluence pages, extract page IDs from URLs
Download JIRA issues (authenticated or public), create background files
Parse HTML, extract JIRA widgets for preservation
Convert markdown to lxml element trees with proper DOM construction
Assemble complete HTML preserving original structure using deepcopy()
Upload to Confluence with error handling
This workflow was developed through multiple failed iterations. The critical lessons learned:
The biggest failure mode: Focusing on technical work while producing shallow content that fails to capture reasoning and trade-offs.
Structured design dialogue that validates ideas before implementation begins.
Comprehensive design intelligence for web and mobile UI/UX across 10 technology stacks.
Comprehensive implementation plans for multi-step tasks, breaking down specs into bite-sized, testable steps.
Introduction to the obra skills system with mandatory skill invocation rules and best practices.
Execute a written implementation plan with critical review and task checkpoints.
Delegate independent tasks to specialized agents working concurrently with isolated context.
Isolated git worktrees with smart directory selection and safety verification.
Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs.
Plan searchable and shareable content that drives traffic, builds authority, and generates leads.
README-first repository scanner that extracts commands and classifies reproduction candidates without executing them.
Brainstorm and prioritize marketing strategies tailored to your SaaS stage, budget, and goals.
Plan and optimize your website's page hierarchy, navigation, URL structure, and internal linking.
The world's fastest calendar for remote work
Capture, organize, and utilize your knowledge effortlessly.
Instant Summaries of Audio & Video Interviews with AnySummary
Transform PDFs into engaging mind maps.
Chat with any PDF instantly
Create an optimal daily plan using your voice
A productivity platform to centralize organizational knowledge and workflows with contextual AI assistance.
AskYourPDF Pricing Plans: Tailored to Your Needs