Use screen -R instead of screen -S and screen -r everywhere

[konard]

Summary

This PR updates all documentation and code to use screen -R instead of screen -S and screen -r for better session management.

Changes Made

Documentation Updates

• README.md: Updated bot session examples from screen -S bot to screen -R bot
• docs/case-studies/issue-986/README.md:
  • Updated all screen command examples to use screen -R
  • Simplified documentation by consolidating create/reattach into single command
  • Updated code examples and usage patterns

Code Changes

• src/start-screen.mjs: Updated console.log messages to suggest screen -R instead of screen -r
• src/telegram-bot.mjs: Updated regex pattern to match screen -R in session name extraction
• src/telegram-solve-queue.lib.mjs: Updated regex pattern to match screen -R in session name extraction
• tests/test-telegram-message-edit-error-handling.mjs: Updated test expectations to match new output format

Experiment Scripts

• experiments/test-issue-356-attach-scenario.sh: Updated comments to reference screen -R
• experiments/test-issue-356-already-attached.sh: Updated attachment command to use screen -R

Changeset

• Added changeset documenting this patch-level change

Why screen -R?

The -R flag is more versatile than -S (create) and -r (reattach) because it:

• Reattaches to existing session if one exists with the given name
• Creates new session if no session exists with that name
• Provides consistent behavior regardless of session state

This makes it the safest and most universal option for session management.

Testing

All changes have been verified:

• Documentation changes are consistent across all files
• Code changes maintain backward compatibility (regex patterns still work)
• Test expectations updated to match new output format
• No breaking changes introduced

Notes

Commands using screen -S <name> -X ... were intentionally NOT changed, as these are used to send commands to existing sessions and the -S flag is appropriate for that use case.

Fixes #1134

---

This PR was created by the AI issue solver

[konard]

🤖 Solution Draft Log

This log file contains the complete execution trace of the AI solution draft process.

💰 Cost estimation:

• Public pricing estimate: $1.508469 USD
• Calculated by Anthropic: $1.228105 USD
• Difference: $-0.280364 (-18.59%)
  📎 Log file uploaded as Gist (634KB)
  🔗 View complete solution draft log

---

Now working session is ended, feel free to review and add any feedback on the solution draft.

[konard]

May be also if we use screen -R we should look in the code fore usages of screen -S (some of which may be not needed anymore).

Get latest version of default branch, make all CI/CD checks pass and ensure all changes are correct, consistent and fully meet the requirements.

[konard]

🤖 AI Work Session Started

Starting automated work session at 2026-01-19T23:04:56.665Z

The PR has been converted to draft mode while work is in progress.

This comment marks the beginning of an AI work session. Please wait working session to finish, and provide your feedback.

[konard]

✅ Work Session Completed

Changes in this session:

1. Merged latest main branch - Incorporated latest changes including the "Fine tune QUEUE_CONFIG" commit
2. Fixed Prettier formatting - Resolved lint errors in changeset file and other files after merge
3. Investigated screen -S usages - Per reviewer request, searched for all screen -S usages. Found that all remaining usages are screen -S <name> -X ... commands which are used to send commands to existing sessions (not to create sessions). These use -S to identify the target session name and are appropriate to keep.
4. Updated test assertions - Tests were failing due to hardcoded QUEUE_CONFIG values that didn't match the new configuration from main. Updated:
   • MIN_START_INTERVAL_MS: 120000 → 60000 (1 minute)
   • RAM_THRESHOLD: 0.5 → 0.65 (65%)
   • CPU_THRESHOLD: 0.5 → 0.75 (75%)
   • CLAUDE_5_HOUR_SESSION_THRESHOLD: 0.9 → 0.85 (85%)
   • CLAUDE_WEEKLY_THRESHOLD: 0.99 → 0.98 (98%)

CI Status:

All checks passing ✅

Summary of PR changes:

• Documentation updated to use screen -R instead of screen -S and screen -r
• Code updated to suggest screen -R in console output messages
• Regex patterns updated to match screen -R in session extraction
• Tests updated to match new output format
• Changeset added for this patch release

---

This work session was performed by the AI issue solver

[konard]

🤖 Solution Draft Log

This log file contains the complete execution trace of the AI solution draft process.

💰 Cost estimation:

• Public pricing estimate: $4.649885 USD
• Calculated by Anthropic: $3.111549 USD
• Difference: $-1.538336 (-33.08%)
  📎 Log file uploaded as Gist (622KB)
  🔗 View complete solution draft log

---

Now working session is ended, feel free to review and add any feedback on the solution draft.