🔍 Documentation Discovery and Task Generation
This is a special initialization task for AI agents to map the project and
create all subsequent tasks.
Your Mission
Investigate this repository to understand:
- What the project does
- What documentation exists
- What's been implemented vs. what's still needed
- What tasks should be created for other agents
Step-by-Step Instructions
1. Explore Project Structure
# Get overview of the repository
find . -type f -name "*.md" | grep -v node_modules | head -20
ls -la docs/ doc/ documentation/ 2>/dev/null
tree -d -L 3 -I 'node_modules|.git|dist|build' 2>/dev/null || find . -type d | head -20
# Check for key files
cat README.md | head -50
cat package.json 2>/dev/null | jq '.name, .description, .scripts'
cat setup.py 2>/dev/null | head -20
2. Identify Documentation
Look for:
- README files at any level
- docs/ or documentation/ directories
- Architecture documents (ARCHITECTURE.md, DESIGN.md)
- API documentation (swagger, openapi files)
- Requirements or specifications
- Development guides (CONTRIBUTING.md, DEVELOPMENT.md)
- TODO files or ROADMAP documents
3. Analyze Implementation Status
# Check source code structure
find src/ -type f -name "*.py" -o -name "*.js" -o -name "*.ts" 2>/dev/null | head -20
find test/ tests/ -type f 2>/dev/null | head -10
# Look for TODO/FIXME comments
grep -r "TODO\|FIXME\|HACK\|BUG" --include="*.py" --include="*.js" \
--include="*.ts" . | head -20
# Check test coverage if available
npm test -- --coverage 2>/dev/null || pytest --cov 2>/dev/null || \
echo "No coverage data"
4. Create Documentation Map
Create .conductor/documentation-map.yaml with this structure:
# Project overview - REQUIRED
project:
name: "[detect from package.json, setup.py, or README]"
description: "[brief description of what this project does]"
type: "[web-app|api|library|cli|mobile|desktop]"
primary_language: "[python|javascript|typescript|go|rust|etc]"
framework: "[react|django|express|etc]"
status: "[prototype|development|beta|production]"
estimated_completion: "[0-100]%"
# Documentation sources - Fill in what exists
documentation:
readme:
- path: "README.md"
summary: "[what this README covers]"
quality: "[excellent|good|needs-work|missing]"
architecture:
- path: "[path to architecture docs]"
summary: "[what it describes]"
decisions: "[list key architectural decisions]"
api:
- path: "[path to API docs]"
format: "[openapi|swagger|markdown|other]"
completeness: "[complete|partial|outdated|missing]"
requirements:
- path: "[path to requirements]"
type: "[functional|technical|business]"
status: "[current|outdated|draft]"
# Current implementation state
implementation:
completed_features:
- name: "[feature name]"
description: "[what it does]"
location: "[where in codebase]"
has_tests: [true|false]
documentation: "[documented|needs-docs|undocumented]"
missing_features:
- name: "[feature from requirements not yet started]"
description: "[what it should do]"
source_requirement: "[where this requirement comes from]"
priority: "[critical|high|medium|low]"
estimated_effort: "[small|medium|large]"
# Proposed tasks - MOST IMPORTANT SECTION
proposed_tasks:
# Create 10-20 specific, actionable tasks based on your investigation
- title: "[Clear, specific task title]"
description: "[What needs to be done]"
type: "[feature|bugfix|refactor|documentation|testing|deployment]"
source_requirement: "[which doc/requirement this comes from]"
estimated_effort: "[small|medium|large]"
priority: "[critical|high|medium|low]"
assigned_role: "[dev|frontend|backend|devops|etc]"
success_criteria:
- "[Specific, measurable criterion]"
- "[Another criterion]"
implementation_notes: "[Any helpful context for the implementer]"
# Summary for humans
summary:
total_tasks: [number]
critical_tasks: [number]
estimated_total_effort: "[in ideal dev days]"
recommended_next_steps:
- "[First thing to do]"
- "[Second thing to do]"
5. Validate Your Work
Before marking complete:
- Ensure the YAML is valid: Run validation with:
python -c "import yaml; yaml.safe_load(open('.conductor/documentation-map.yaml'))"
- Check you've created at least 10 concrete tasks
- Verify each task has clear success criteria
- Make sure priorities are reasonable
Success Criteria
Completion
After creating the documentation map:
- Run the task generator:
python .conductor/scripts/generate-tasks-from-map.py --auto
- Verify tasks were created:
gh issue list -l 'conductor:task' --limit 25
- Comment on this issue with a summary of tasks created
- Mark this task complete using:
./conductor complete
This is a one-time initialization task. Once complete, all future work will be
properly coordinated.
🔍 Documentation Discovery and Task Generation
This is a special initialization task for AI agents to map the project and
create all subsequent tasks.
Your Mission
Investigate this repository to understand:
Step-by-Step Instructions
1. Explore Project Structure
2. Identify Documentation
Look for:
3. Analyze Implementation Status
4. Create Documentation Map
Create
.conductor/documentation-map.yamlwith this structure:5. Validate Your Work
Before marking complete:
python -c "import yaml; yaml.safe_load(open('.conductor/documentation-map.yaml'))"Success Criteria
.conductor/documentation-map.yamlCompletion
After creating the documentation map:
python .conductor/scripts/generate-tasks-from-map.py --autogh issue list -l 'conductor:task' --limit 25./conductor completeThis is a one-time initialization task. Once complete, all future work will be
properly coordinated.