feat: add automated skill evaluation workflow and reporting

[yongsinp]

Overview

This pull request introduces a comprehensive automated evaluation workflow for skills, including new scripts and a GitHub Actions workflow for continuous skill review and test case-based evaluation. These changes enable automatic skill discovery, review, evaluation, and summary reporting, with support for test case generation, evaluation using multiple models and detailed Markdown reports.

Changes

New Evaluation Workflow Automation

Added .github/workflows/evaluate-all-skills.yml, a new GitHub Actions workflow that:

• Discovers all skills and identifies those with evaluation test cases (retrieved from private rse-plugins-testcases).
• Reviews each skill using the tessl tool and extracts validation and LLM judge results.
• Evaluates skills with test cases using skill-eval-action across all available models.
• Generates a Markdown summary report using a new evaluation summary script.

Test Case Generation

Added .github/scripts/generate-testcases.py, a script that automatically generates 5 to 6 YAML evaluation test cases when a new skill gets added, including both positive and negative trigger scenarios, based on the contents of each SKILL.md.

Evaluation Summary Rendering

Added .github/scripts/render-eval-summary.py, a script that combines review and evaluation JSON results into a detailed Markdown report with per-skill breakdowns, pass rates, and model-by-model comparisons for GitHub Actions reporting.

Skill Updates

Updated skills based on feedback from the Tessl LLM judge.