v0.7.1

• tech-book: New skill — generate complete technical books with mdbook/pandoc, deploy to GitHub Pages
• md2pdf: Removed — functionality fully covered by any2pdf (reportlab + pandoc fallback)

v0.7.0

• skill-optimizer: Add Step 7 — auto commit & push after optimization (no more manual git push)
• finder-action: Fix sandbox entitlements — use temporary-exception.files.absolute-path.read-write instead of files.user-selected.read-write, document sandbox-must-be-on requirement

v0.6.1

Release v0.6.1

v0.6.0

• proposal: New skill — generate complete business proposals from client requirements (architecture, budget, timeline, risk analysis)
• thesis-polish: New skill — MBA thesis polishing to national outstanding thesis standards
• finder-action: New skill — generate Mac Finder right-click menu actions (Quick Action or Finder Sync Extension)
• pdf2png: New skill — PDF → single concatenated PNG via macOS CoreGraphics
• md2pdf: New skill — Markdown → PDF via pandoc + xelatex with CJK support
• gh-tidy: New skill — interactive GitHub repo triage (issues, PRs, stale branches, labels)
• obsidian-reset-cache: New skill — reset Obsidian cache to fix "Loading cache..." hang
• any2pdf: Add pandoc + XeLaTeX fallback engine, latex-clean theme, watermark params, image-cover mode
• README: Auto-generate skills table from SKILL.md frontmatter via CI, shield badges for stats
• CI: Add sync-readme.yml workflow to keep skills table in sync on push

v0.5.2

Release v0.5.2

v0.5.1

• README: Reorganize skills into 6 categories (Meta Skills / Document Conversion / Content Processing / Content Creation / xBTI / Dev Tools), add missing document-illustrator skill, extract Theme Gallery to docs/THEME-GALLERY.md with external links from skill descriptions

v0.5.0

• document-illustrator: New skill (migrated from op7418/Document-illustrator-skill). Redesigned workflow: backup → global insertion planning → parallel image generation → async insert-in-place → cleanup. Uses anchor-text positioning instead of line numbers to handle offset drift. All images generated concurrently via Agent tool. SKILL.md trimmed from 480 to ~120 lines.

v0.4.1

• any2pdf: Fix silent image drops. Relative ![alt](path) references are now resolved against the input markdown's directory (not cwd), missing images emit a WARN: image not found to stderr instead of being silently dropped, and multi-line image refs (caused by pandoc's default --wrap=auto) are collapsed during preprocessing. SKILL.md gains an "Input Format" section clarifying markdown-only input and a pandoc --wrap=none pipeline tip. Bumps skill version 1.0.0 → 1.0.1.

v0.4.0

• skill-optimizer: New meta skill — audits an existing skill against repo conventions + official skill-creator best practices, applies fixes, bumps semver version, and prepends a per-skill CHANGELOG.md entry. Fully automatic, context-driven (prioritizes issues raised in the current conversation). Ships lint_skill.py + bump_version.py (stdlib only).

v0.3.2

• xbti-creator: Fix step numbering (duplicate Step 2), fix Gallery PR submission (wrong gh fork syntax, missing index.js/case.json/registry updates, image→images dir), add Gallery opt-in to initial questionnaire