v1.64.0: claude-usage-analyst, bilibili-source + skill updates

Cumulative release since v1.60.1. Catalog now at 63 skills, metadata 1.64.0.

✨ New Skills

• claude-usage-analyst — turns local ccusage data into a human-readable explanation of Claude Code / Claude Desktop token usage, cost, quota burn, and cache pressure (suite: daymade-claude-code).
• bilibili-source — login-free fetch of comprehensive Bilibili (B站) video data (stats, UP fans, tags, per-part cids, full danmaku) in one view/detail call.

🔧 Skill Updates

• skill-creator v1.2.0 — five incident-distilled authoring rules + a "Phase 9 实战案例库"; quick_validate self-fix.
• debugging-network-issues v1.2.0 — Cloudflare-524 upload-timeout case study, Step 0.6 upload-vs-processing recipe, cognitive Traps 10 & 12.
• macos-cleaner v1.2.0 — progressive-disclosure refactor + safety guidance aligned to "user runs cleanup, never auto-execute" (thanks @geniusart, #84).
• tunnel-doctor v1.6.0 — TUN measurement-contamination diagnostics.
• pdf-creator — new `warm-terra-menu` theme for 2-column long-text menus.

🐛 Fixes

• SKILL.md frontmatter strict-YAML validity (codex compatibility) — tunnel-doctor, benchmark-due-diligence, pdf-creator.

📖 Full changelog: https://github.com/daymade/claude-code-skills/blob/main/CHANGELOG.md

Release v1.60.1: Harden macos-cleaner safety

Fixed

• Hardened macos-cleaner safe deletion by blocking high-risk system and credential paths before confirmation and inside delete_path().
• Updated app remnant detection to read installed app Bundle Identifiers and avoid misclassifying installed apps as orphaned data.
• Added regression coverage for blocked paths, malformed app plists, Bundle Identifier matching, and legacy app-name matching.

Fixes #70.

Verification

• python3 -m unittest tests/test_macos_cleaner_safety.py
• python3 -m py_compile macos-cleaner/scripts/safe_delete.py macos-cleaner/scripts/find_app_remnants.py tests/test_macos_cleaner_safety.py
• uv run --with PyYAML python -m scripts.quick_validate ../../macos-cleaner
• uv run --with PyYAML python -m scripts.security_scan ../../macos-cleaner --verbose
• bash daymade-claude-code/marketplace-dev/scripts/check_marketplace.sh
• python3 macos-cleaner/scripts/safe_delete.py /System
• python3 macos-cleaner/scripts/safe_delete.py ~/.ssh

pdf-creator v1.3.1: Complete CJK coverage + pandoc-highlighted code blocks

What's Changed

Follow-up patch to v1.3.0 addressing two incomplete gaps in the CJK code block fix.

Bug Fixes

• Japanese/Korean no longer garbled: v1.3.0 only covered Chinese (CJK Unified Ideographs). Hiragana, Katakana, and Hangul code blocks now render correctly.
• Syntax-highlighted code blocks now handled: Pandoc wraps language-tagged blocks as <pre class="sourceCode python"><code class="sourceCode python">. v1.3.0's regex only matched plain <pre><code>, so \``pythonblocks with Chinese comments were skipped. Relaxed regex + strip` highlight wrappers inside converted CJK blocks.

Quality

• Adds scripts/tests/test_cjk_code_blocks.py regression test (6 scenarios: Chinese, Japanese, pure-ASCII, inline code, PDF generation, extracted text integrity)
• Fixes README.md quickstart/requirements still referencing the obsolete markdown Python package (the script uses pandoc via subprocess)

Upgrade

claude plugin marketplace add https://github.com/daymade/claude-code-skills
claude plugin install pdf-creator

pdf-creator v1.3.0: Fix CJK code block rendering

What's Changed

Fix CJK text garbled in weasyprint code blocks

weasyprint renders <pre> blocks with monospace fonts that lack CJK glyphs, causing Chinese/Japanese/Korean characters to display as garbled text.

Changes

• Add _fix_cjk_code_blocks() preprocessor: detects CJK in <pre><code> → <div class="cjk-code-block"> with inherited body font
• Pure-ASCII code blocks left untouched
• Add code/pre/pre code CSS to both themes (previously missing)
• Document CJK troubleshooting in SKILL.md

Upgrade

claude plugin marketplace add https://github.com/daymade/claude-code-skills
claude plugin install pdf-creator

Release v1.37.0: Add excel-automation and capture-screen

New Skills

excel-automation

Create formatted Excel workbooks, parse complex .xlsm files via stdlib XML/ZIP workflows, and control Excel windows on macOS via AppleScript.

Installation:

claude plugin install excel-automation@daymade-skills

capture-screen

Programmatic macOS screenshot capture workflows with Swift CGWindowID discovery and screencapture window capture patterns.

Installation:

claude plugin install capture-screen@daymade-skills

Improvements

• Fixed missing promptfoo-evaluation/scripts/metrics.py referenced by skill examples
• Bumped promptfoo-evaluation plugin version to 1.1.0
• Updated marketplace metadata to 1.37.0 and synchronized docs:
  • README.md
  • README.zh-CN.md
  • CLAUDE.md
  • CHANGELOG.md

Marketplace Summary

• Skills/plugins: 41
• Marketplace version: 1.37.0

v1.33.0: tunnel-doctor v1.1.0 — SSH Tunnel SOP & Makefile Patterns

tunnel-doctor v1.0.0 → v1.1.0

New: Remote Development SOP

Proactive setup guide for remote development over Tailscale with proxy tools:

• Proxy-safe Makefile pattern — --noproxy localhost for all health check and warmup curl commands
• SSH tunnel Makefile targets — tunnel (foreground) and tunnel-bg (background with autossh auto-reconnect)
• Multi-port tunnel configuration — forward dev server + S3 + API gateway in one tunnel
• SSH non-login shell setup — source ~/.zshrc pattern with Makefile REMOTE_CMD macro
• End-to-end workflow — first-time setup + daily development flow
• Pre-flight checklist — 10 verification items before starting remote development

New Diagnostic Steps

• Step 2D: Auth redirect fix via SSH local port forwarding (solves APP_URL=localhost redirect issue)
• Step 2E: localhost proxy interception in Makefiles and scripts

Quality Improvements (via skill-reviewer)

• Fixed step ordering: 2A→2B→2C→2D→2E (was 2A→2C→2D→2E→2B)
• Fixed description to third-person voice per skill best practices
• Replaced 5 hardcoded IPs with <tailscale-ip> placeholder
• Deduplicated SSH non-login shell content (references proxy_fixes.md)
• New keywords: ssh-tunnel, autossh, makefile, remote-development

References Updated

• Added SSH non-login shell pitfall section
• Added localhost proxy interception section
• Strengthened --data-binary vs -d warning (prevents config destruction)

Release v1.32.1: Strengthen claude-md-progressive-disclosurer anti-deletion safeguards

Updated Skill: claude-md-progressive-disclosurer v1.2.0

Strengthened safeguards against information loss during CLAUDE.md optimization.

What's New

• "Move as-is, no compression" iron rule — Content moved to Level 2 must be copied verbatim, no "incidental trimming"
• Anti-pattern 6: Compression during move — Moving and simplifying are two separate operations, never combine them
• Anti-pattern 7: Disguising loss as "intentional deletion" — Every deletion must have a pre-decision with canonical source, not post-hoc rationalization
• Enhanced Step 5 verification — Three mandatory sub-checks: file existence (5a), content completeness with diff script (5b), no line counting (5c)
• Ban wc -l globally — Prohibited at every workflow stage, including final summary
• Real-world case studies 8 & 9 — Documented actual incidents of info loss from compression-during-move and post-hoc deletion classification

Installation

claude plugin install claude-md-progressive-disclosurer@daymade-skills

🤖 Generated with Claude Code

Release v1.30.0: Add competitors-analysis skill

New Skill: competitors-analysis

Evidence-based competitor tracking and analysis. All analysis must be based on actual code, never assumptions.

Features

• Pre-analysis checklist to ensure repositories are cloned locally before any analysis
• Forbidden patterns to prevent assumptions ("推测...", "可能...", "应该...")
• Required patterns for source citation (file:line_number format)
• Tech stack analysis guides for Node.js, Python, Rust projects
• Directory structure conventions for organized competitor tracking
• Bundled templates: profile_template.md, analysis_checklist.md
• Management script: update-competitors.sh (clone/pull/status)

Installation

claude plugin marketplace update daymade-skills
claude plugin install competitors-analysis@daymade-skills

Usage

# Then ask Claude to analyze a competitor
"分析竞品 https://github.com/org/repo"
"添加竞品到 flowzero 产品的竞品列表"

Full Changelog

See CHANGELOG.md for complete changes.

---

🤖 Generated with Claude Code

Release v1.29.0: Enhance skill-creator with YAML frontmatter documentation

Enhanced Skill: skill-creator v1.4.0

This release significantly improves skill-creator documentation based on community contribution (PR #6 by @costa-marcello).

New Features

• Complete YAML Frontmatter Reference: Comprehensive table documenting all available fields (name, description, context, agent, disable-model-invocation, user-invocable, allowed-tools, model, argument-hint, hooks)

• context: fork Documentation: Critical guidance for creating skills accessible to subagents via Task tool

• Invocation Control Table: Clear comparison showing how different frontmatter combinations affect skill accessibility (user invocable, Claude invocable, subagent accessible)

• $ARGUMENTS Placeholder: Documented how skill content can receive user input via /skill-name arguments

• Enhanced Examples:

  • Wildcard syntax for allowed-tools: Bash(git *), Bash(npm *), Bash(docker compose *)
  • Inline hooks configuration example
  • Task-based skill with subagent execution
  • Reference skill that runs inline

• Updated init_skill.py Template: New skills now include commented optional frontmatter fields

Installation

claude plugin marketplace update daymade-skills
claude plugin install skill-creator@daymade-skills

Contributors

• @costa-marcello - PR #6: Initial frontmatter documentation

---

🤖 Generated with Claude Code

Release v1.28.0: Enhance meeting-minutes-taker with speaker identification

Enhanced Skill: meeting-minutes-taker v1.1.0

Add speaker identification and pre-processing pipeline for better transcript processing.

New Features

Speaker Identification

Analyze transcript features to identify anonymous speakers:

Feature	What to Look For
Word count	Total words (high = senior/lead, low = observer)
Segment count	Speaking frequency (frequent = active participant)
Avg segment length	Words per turn (long = presenter, short = responder)
Filler ratio	% of filler words - low = prepared speaker

Context File Template

New references/context_file_template.md for team directory mapping:

• Team members with roles and communication styles
• Project terminology and naming conventions
• Automatic speaker matching based on features

Intelligent File Naming

Auto-generate output filename: YYYY-MM-DD-<topic>-<type>.md

Examples:

• 2026-01-25-order-api-design-review.md
• 2026-01-20-sprint-planning.md

Pre-processing Pipeline

Integrated with other skills:

• markdown-tools - Convert .docx/.pdf to Markdown first
• transcript-fixer - Fix ASR/STT errors before processing

Full Changelog

https://github.com/daymade/claude-code-skills/blob/main/CHANGELOG.md