Implement learnings persistence in Ralph Loop workflow

[Copilot]

Adds iterative learning workflow that accumulates knowledge across iterations via append-only progress log and pattern repository.

Changes

Workflow: .github/workflows/ralph-loop-basic.md

• Iterative execution with configurable iteration count (default: 3)
• Pre-iteration learning context load from progress.txt and AGENTS.md
• Post-iteration structured append to progress.txt:

  ---
  2026-01-24 06:00:00 UTC - Iteration 2/3 - Run 12345678
  
  Task: [description]
  
  ## Work Completed
  ## Key Learnings
  ## Gotchas
  ## Next Steps
  ---

• Selective AGENTS.md updates for reusable patterns (code patterns, gotchas, test strategies)
• Local git commits per iteration, single PR creation on completion

Documentation: examples/ralph/README.md

• Usage patterns and format specifications
• Guidelines for iteration count selection and learning documentation
• Troubleshooting common issues (missing updates, unapplied learnings)

Design Notes

• Uses workflow_dispatch with task (required) and iterations (optional) inputs
• Read-only permissions; PR creation via safe-outputs.create-pull-request
• Progress log is append-only; AGENTS.md filters task-specific details
• Follows existing patterns from changeset.md and daily-doc-updater.md

Original prompt

---

This section details on the original issue you should resolve

<issue_title>[plan] Implement learnings persistence in Ralph Loop workflow</issue_title>
<issue_description>## Objective

Enhance the Ralph Loop workflow to persist learnings between iterations, similar to Ralph's progress.txt and AGENTS.md updates.

Context

A key feature of Ralph is that learnings from each iteration persist and inform future iterations. This includes both progress.txt (append-only log) and AGENTS.md updates (discovered patterns and conventions).

Approach

1. Update .github/workflows/ralph-loop-basic.md to:

   • Append learnings to progress.txt after each iteration
   • Extract key learnings from AI responses
   • Update relevant AGENTS.md files with discovered patterns
   • Commit progress.txt and AGENTS.md changes

2. Add logic to extract learnings:

   • Code patterns discovered
   • Gotchas encountered
   • Useful conventions
   • Test patterns

3. Format progress.txt entries with timestamp and iteration info

Files to Modify

• .github/workflows/ralph-loop-basic.md - Add learnings persistence
• examples/ralph/README.md - Document learnings feature

Acceptance Criteria

• Workflow appends to progress.txt after each iteration
• Each entry includes timestamp and iteration number
• Workflow updates AGENTS.md with discovered patterns
• Learnings are extractable and well-formatted
• Git commits include both progress.txt and AGENTS.md changes
  Related to Add Ralph Loop examples for agentic workflows #11132

AI generated by Plan Command for #11132

Comments on the Issue (you are @copilot in this section)

• Fixes [plan] Implement learnings persistence in Ralph Loop workflow #11137

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.