Improve SkillOpt UI progress and review flow

[JamesMaki]

Summary

• place each SkillOpt run log directly beneath the skill being optimized
• show a few-minutes notice, live log updates, and per-skill version counts
• make Review change navigate to the evidence panel
• simplify promotion to the review-card click plus one final modal action
• pass the generic API_KEY into OpenRouter reflection calls
• document the version explorer and refresh all UI screenshots
• link the SkillOpt and Langfuse GitHub repositories from the README

Verification

• docker run --rm -v "$PWD:/app" -w /app ingot-mcp python -m pytest tests/test_ui.py tests/test_optimize.py tests/test_env_check.py -q (153 passed)
• final tests/test_ui.py run after version counts (77 passed)
• live OpenRouter reflection smoke request succeeded
• rebuilt and restarted the local UI at http://localhost:8080
• commit is GPG signed

CodeScene was not run because it was not explicitly requested.

Summary by CodeRabbit

• New Features

  • Added inline SkillOpt activity logs for each skill, with automatic scrolling during optimization.
  • Added version and rollback status details to the skills library.
  • Streamlined the review flow with a direct Approve & promote action.
  • Improved pending-review navigation and focus behavior.

• Bug Fixes

  • Improved OpenRouter authentication during optimization.

• Documentation

  • Expanded guidance for SkillOpt optimization, review, promotion, and read-only version exploration.
  • Updated UI screenshots and Langfuse setup information.

[coderabbitai]

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 12ce22ad-2551-4095-a116-1ce5d982e351

📥 Commits

Reviewing files that changed from the base of the PR and between bda5fc5 and 1f395bd.

📒 Files selected for processing (10)

• README.md
• docs/how-it-works.md
• docs/tutorial.md
• docs/ui-compare.webp
• docs/ui-home.webp
• docs/ui-review.webp
• optimize/rollout.py
• tests/test_optimize.py
• tests/test_ui.py
• ui/static/index.html

---

📝 Walkthrough

Walkthrough

The UI now displays SkillOpt activity within each skill, uses a direct approval-and-promotion action, supports focused pending-review navigation, explicitly forwards OpenRouter credentials, and updates documentation and screenshots for the revised workflow.

Changes

SkillOpt control flow

Layer / File(s)	Summary
Inline SkillOpt rendering ui/static/index.html, tests/test_ui.py	Client state and refresh logic now pass revision history and run data into per-skill rendering, which shows version counts and nested SkillOpt logs instead of a standalone activity panel.
Review navigation and promotion ui/static/index.html, tests/test_ui.py	Pending review navigation can scroll and focus the review section, while Approve & promote directly executes promotion without a separate confirmation control.
OpenRouter reflection credentials optimize/rollout.py, tests/test_optimize.py	OpenRouter reflection calls explicitly pass the generic API key, with coverage for the model, base URL, key, and returned completion.
UI workflow documentation README.md, docs/how-it-works.md, docs/tutorial.md, docs/*.webp	Documentation and screenshots describe the read-only explorer, SkillOpt progress logs, revised approval flow, and Langfuse setup.

Estimated code review effort: 3 (Moderate) | ~20 minutes

Possibly related PRs

• SlanchaAI/ingot#21: Modifies the same UI approval path with WebAuthn and promotion request handling.
• SlanchaAI/ingot#23: Modifies the same OpenRouter reflection/teacher model wiring.
• SlanchaAI/ingot#34: Modifies the same UI refresh, polling, and per-skill run-log rendering flow.

Suggested reviewers: laulpogan

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch agent/inline-skillopt-progress

---

_{Comment @coderabbitai help to get the list of available commands.}