/queue-learnings

Review the current work context and recent activities to identify any learnings or remaining tasks that still require work but are not captured in existing job files.

Process

Step 1: Scan Existing Jobs

Check what jobs are already scheduled:

• List all files in jobs/ directory
• Read job titles and descriptions to understand what's already covered
• Note any gaps in the planned work

Step 2: Analyze Recent Context

Review recent work to identify uncaptured learnings:

• Consider any issues discovered during recent implementations
• Look for incomplete solutions or workarounds that need proper fixes
• Look for errors we encountered, and ways we could avoid them in the future
• Identify any best practices or patterns that emerged
• Check for any technical debt or follow-up items mentioned

Step 3: Identify Missing Work

Ask the key question: "Did we learn anything that still requires work that is not captured in jobs already scheduled in jobs/*.md?"

Consider categories like:

• Bug fixes needed
• Performance improvements
• Code quality enhancements
• Documentation updates
• Testing gaps
• Refactoring opportunities
• New features identified during implementation
• Integration improvements
• User experience enhancements
• Errors we encounterd
• Problems we had to solve that could be resolved once and for all

Step 4: Create Jobs for Missing Items

If any missing work is identified, use the addjob agent to create comprehensive job files that are self-sufficient and self-documenting. Each job should include:

Essential Context:

• Complete background: What project/work led to this discovery
• Root cause: What problem occurred that revealed this need
• Current state: What exists now and what's missing
• Impact: Why this work matters and what it improves

Detailed Problem Description:

• Specific symptoms or issues encountered
• Examples of when/where the problem manifests
• Current workarounds being used (if any)
• How the problem was discovered

Implementation Roadmap:

• Step-by-step approach to solve the problem
• Specific files, directories, or components involved
• Technical requirements and dependencies
• Expected deliverables and success criteria

Reference Information:

• Links to related files or directories
• Commands or tools that need to be used
• Examples of similar implementations
• Any research or investigation already done

Use the Task tool to invoke the addjob agent with COMPLETE context:

Task(subagent_type: 'addjob', description: 'Create comprehensive self-documenting jobs', prompt: 'Create detailed job files for the following work items. Each job must be completely self-sufficient with full background context, problem description, implementation details, and all reference information needed to execute the work without requiring additional context from previous conversations: [provide exhaustive details for each item]')

Step 5: Report Results

Summarize what was found:

• List any new jobs created
• Confirm if all learnings are now captured
• Note if no additional work was identified

Focus on being thorough but practical - capture genuine work items that add value, not busy work.

Job Quality Standards

Each created job must pass the "6 months later test" - someone (including yourself) should be able to execute the job 6 months from now without needing to:

• Remember the conversation that created it
• Hunt for missing context or background information
• Guess what files or components are involved
• Figure out why the work is needed

Good job example:

• "During git rewrite-commit-descriptions development, we encountered bash syntax ${var,,} that failed on macOS default bash 3.2. This broke the detect-bad-commits.sh script. Need to establish coding standards..."

Bad job example:

• "Fix bash compatibility issues" (no context about what, where, or why)